generic-library/opencv
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
opencv/tests/cxcore/src/amath.cpp

3353 lines
104 KiB
C++
Raw Blame History

/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// Intel License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000, Intel Corporation, all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of Intel Corporation may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
//////////////////////////////////////////////////////////////////////////////////////////
/////////////////// tests for matrix operations and math functions ///////////////////////
//////////////////////////////////////////////////////////////////////////////////////////
#include "cxcoretest.h"
#include <float.h>
#include <math.h>
/// !!! NOTE !!! These tests happily avoid overflow cases & out-of-range arguments
/// so that output arrays contain neigher Inf's nor Nan's.
/// Handling such cases would require special modification of check function
/// (validate_test_results) => TBD.
/// Also, need some logarithmic-scale generation of input data. Right now it is done (in some tests)
/// by generating min/max boundaries for random data in logarimithic scale, but
/// within the same test case all the input array elements are of the same order.
static const CvSize math_sizes[] = {{10,1}, {100,1}, {10000,1}, {-1,-1}};
static const int math_depths[] = { CV_32F, CV_64F, -1 };
static const char* math_param_names[] = { "size", "depth", 0 };
static const CvSize matrix_sizes[] = {{3,3}, {4,4}, {10,10}, {30,30}, {100,100}, {500,500}, {-1,-1}};
class CxCore_MathTestImpl : public CvArrTest
{
public:
CxCore_MathTestImpl( const char* test_name, const char* test_funcs );
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
double get_success_error_level( int /*test_case_idx*/, int i, int j );
bool test_nd;
};
CxCore_MathTestImpl::CxCore_MathTestImpl( const char* test_name, const char* test_funcs )
: CvArrTest( test_name, test_funcs, "" )
{
optional_mask = false;
test_array[INPUT].push(NULL);
test_array[OUTPUT].push(NULL);
test_array[REF_OUTPUT].push(NULL);
default_timing_param_names = math_param_names;
size_list = math_sizes;
whole_size_list = 0;
depth_list = math_depths;
cn_list = 0;
test_nd = false;
}
double CxCore_MathTestImpl::get_success_error_level( int /*test_case_idx*/, int i, int j )
{
return CV_MAT_DEPTH(test_mat[i][j].type) == CV_32F ? FLT_EPSILON*128 : DBL_EPSILON*1024;
}
void CxCore_MathTestImpl::get_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
int depth = cvTsRandInt(rng)%2 + CV_32F;
int cn = cvTsRandInt(rng) % 4 + 1, type = CV_MAKETYPE(depth, cn);
int i, j;
CvArrTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
for( i = 0; i < max_arr; i++ )
{
int count = test_array[i].size();
for( j = 0; j < count; j++ )
types[i][j] = type;
}
test_nd = cvTsRandInt(rng)%3 == 0;
}
CxCore_MathTestImpl math_test( "math", "" );
class CxCore_MathTest : public CxCore_MathTestImpl
{
public:
CxCore_MathTest( const char* test_name, const char* test_funcs );
};
CxCore_MathTest::CxCore_MathTest( const char* test_name, const char* test_funcs )
: CxCore_MathTestImpl( test_name, test_funcs )
{
size_list = 0;
depth_list = 0;
}
////////// exp /////////////
class CxCore_ExpTest : public CxCore_MathTest
{
public:
CxCore_ExpTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void get_minmax_bounds( int i, int j, int type, CvScalar* low, CvScalar* high );
double get_success_error_level( int /*test_case_idx*/, int i, int j );
int prepare_test_case( int test_case );
void run_func();
void prepare_to_validation( int test_case_idx );
int out_type;
};
CxCore_ExpTest::CxCore_ExpTest()
: CxCore_MathTest( "math-exp", "cvExp" )
{
out_type = 0;
}
double CxCore_ExpTest::get_success_error_level( int /*test_case_idx*/, int /*i*/, int /*j*/ )
{
int in_depth = CV_MAT_DEPTH(test_mat[INPUT][0].type);
int out_depth = CV_MAT_DEPTH(test_mat[OUTPUT][0].type);
int min_depth = MIN(in_depth, out_depth);
return min_depth == CV_32F ? 1e-5 : 1e-8;
}
void CxCore_ExpTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CxCore_MathTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
out_type = types[OUTPUT][0];
/*if( CV_MAT_DEPTH(types[INPUT][0]) == CV_32F && (cvRandInt(ts->get_rng()) & 3) == 0 )
types[OUTPUT][0] = types[REF_OUTPUT][0] =
out_type = (types[INPUT][0] & ~CV_MAT_DEPTH_MASK)|CV_64F;*/
}
void CxCore_ExpTest::get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high )
{
double l = cvTsRandReal(ts->get_rng())*10+1;
double u = cvTsRandReal(ts->get_rng())*10+1;
l *= -l;
u *= u;
*low = cvScalarAll(l);
*high = cvScalarAll(CV_MAT_DEPTH(out_type)==CV_64F? u : u*0.5);
}
int CxCore_ExpTest::prepare_test_case( int test_case )
{
int code = CxCore_MathTest::prepare_test_case(test_case);
if( code < 0 )
return code;
CvRNG* rng = ts->get_rng();
int i, j, k, count = cvTsRandInt(rng) % 10;
CvMat* src = &test_mat[INPUT][0];
int depth = CV_MAT_DEPTH(src->type);
// add some extremal values
for( k = 0; k < count; k++ )
{
i = cvTsRandInt(rng) % src->rows;
j = cvTsRandInt(rng) % (src->cols*CV_MAT_CN(src->type));
int sign = cvTsRandInt(rng) % 2 ? 1 : -1;
if( depth == CV_32F )
((float*)(src->data.ptr + src->step*i))[j] = FLT_MAX*sign;
else
((double*)(src->data.ptr + src->step*i))[j] = DBL_MAX*sign;
}
return code;
}
void CxCore_ExpTest::run_func()
{
if(!test_nd)
cvExp( test_array[INPUT][0], test_array[OUTPUT][0] );
else
{
cv::MatND a = cv::cvarrToMatND(test_array[INPUT][0]);
cv::MatND b = cv::cvarrToMatND(test_array[OUTPUT][0]);
cv::exp(a, b);
}
}
void CxCore_ExpTest::prepare_to_validation( int /*test_case_idx*/ )
{
CvMat* a = &test_mat[INPUT][0];
CvMat* b = &test_mat[REF_OUTPUT][0];
int a_depth = CV_MAT_DEPTH(a->type);
int b_depth = CV_MAT_DEPTH(b->type);
int ncols = test_mat[INPUT][0].cols*CV_MAT_CN(a->type);
int i, j;
for( i = 0; i < a->rows; i++ )
{
uchar* a_data = a->data.ptr + i*a->step;
uchar* b_data = b->data.ptr + i*b->step;
if( a_depth == CV_32F && b_depth == CV_32F )
{
for( j = 0; j < ncols; j++ )
((float*)b_data)[j] = (float)exp((double)((float*)a_data)[j]);
}
else if( a_depth == CV_32F && b_depth == CV_64F )
{
for( j = 0; j < ncols; j++ )
((double*)b_data)[j] = exp((double)((float*)a_data)[j]);
}
else
{
assert( a_depth == CV_64F && b_depth == CV_64F );
for( j = 0; j < ncols; j++ )
((double*)b_data)[j] = exp(((double*)a_data)[j]);
}
}
}
CxCore_ExpTest exp_test;
////////// log /////////////
class CxCore_LogTest : public CxCore_MathTest
{
public:
CxCore_LogTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void get_minmax_bounds( int i, int j, int type, CvScalar* low, CvScalar* high );
void run_func();
void prepare_to_validation( int test_case_idx );
};
CxCore_LogTest::CxCore_LogTest()
: CxCore_MathTest( "math-log", "cvLog" )
{
}
void CxCore_LogTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CxCore_MathTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
/*if( CV_MAT_DEPTH(types[INPUT][0]) == CV_32F && (cvRandInt(ts->get_rng()) & 3) == 0 )
types[INPUT][0] = (types[INPUT][0] & ~CV_MAT_DEPTH_MASK)|CV_64F;*/
}
void CxCore_LogTest::get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high )
{
double l = cvTsRandReal(ts->get_rng())*15-5;
double u = cvTsRandReal(ts->get_rng())*15-5;
double t;
l = exp(l);
u = exp(u);
if( l > u )
CV_SWAP( l, u, t );
*low = cvScalarAll(l);
*high = cvScalarAll(u);
}
void CxCore_LogTest::run_func()
{
if(!test_nd)
cvLog( test_array[INPUT][0], test_array[OUTPUT][0] );
else
{
cv::MatND a = cv::cvarrToMatND(test_array[INPUT][0]);
cv::MatND b = cv::cvarrToMatND(test_array[OUTPUT][0]);
cv::log(a, b);
}
}
void CxCore_LogTest::prepare_to_validation( int /*test_case_idx*/ )
{
CvMat* a = &test_mat[INPUT][0];
CvMat* b = &test_mat[REF_OUTPUT][0];
int a_depth = CV_MAT_DEPTH(a->type);
int b_depth = CV_MAT_DEPTH(b->type);
int ncols = test_mat[INPUT][0].cols*CV_MAT_CN(a->type);
int i, j;
for( i = 0; i < a->rows; i++ )
{
uchar* a_data = a->data.ptr + i*a->step;
uchar* b_data = b->data.ptr + i*b->step;
if( a_depth == CV_32F && b_depth == CV_32F )
{
for( j = 0; j < ncols; j++ )
((float*)b_data)[j] = (float)log((double)((float*)a_data)[j]);
}
else if( a_depth == CV_64F && b_depth == CV_32F )
{
for( j = 0; j < ncols; j++ )
((float*)b_data)[j] = (float)log(((double*)a_data)[j]);
}
else
{
assert( a_depth == CV_64F && b_depth == CV_64F );
for( j = 0; j < ncols; j++ )
((double*)b_data)[j] = log(((double*)a_data)[j]);
}
}
}
CxCore_LogTest log_test;
////////// pow /////////////
static const double math_pow_values[] = { 2., 5., 0.5, -0.5, 1./3, -1./3, CV_PI };
static const char* math_pow_param_names[] = { "size", "power", "depth", 0 };
static const int math_pow_depths[] = { CV_8U, CV_16U, CV_16S, CV_32S, CV_32F, CV_64F, -1 };
class CxCore_PowTest : public CxCore_MathTest
{
public:
CxCore_PowTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void get_minmax_bounds( int i, int j, int type, CvScalar* low, CvScalar* high );
void get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images );
int write_default_params( CvFileStorage* fs );
void print_timing_params( int test_case_idx, char* ptr, int params_left );
void run_func();
int prepare_test_case( int test_case_idx );
void prepare_to_validation( int test_case_idx );
double get_success_error_level( int test_case_idx, int i, int j );
double power;
};
CxCore_PowTest::CxCore_PowTest()
: CxCore_MathTest( "math-pow", "cvPow" )
{
power = 0;
default_timing_param_names = math_pow_param_names;
depth_list = math_pow_depths;
}
void CxCore_PowTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
int depth = cvTsRandInt(rng) % (CV_64F+1);
int cn = cvTsRandInt(rng) % 4 + 1;
int i, j;
CvArrTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
depth += depth == CV_8S;
if( depth < CV_32F || cvTsRandInt(rng)%8 == 0 )
// integer power
power = (int)(cvTsRandInt(rng)%21 - 10);
else
{
i = cvTsRandInt(rng)%17;
power = i == 16 ? 1./3 : i == 15 ? 0.5 : i == 14 ? -0.5 : cvTsRandReal(rng)*10 - 5;
}
for( i = 0; i < max_arr; i++ )
{
int count = test_array[i].size();
int type = CV_MAKETYPE(depth, cn);
for( j = 0; j < count; j++ )
types[i][j] = type;
}
test_nd = cvTsRandInt(rng)%3 == 0;
}
void CxCore_PowTest::get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images )
{
CxCore_MathTest::get_timing_test_array_types_and_sizes( test_case_idx,
sizes, types, whole_sizes, are_images );
power = cvReadReal( find_timing_param( "power" ), 0.2 );
}
int CxCore_PowTest::write_default_params( CvFileStorage* fs )
{
int i, code = CxCore_MathTest::write_default_params(fs);
if( code < 0 || ts->get_testing_mode() != CvTS::TIMING_MODE )
return code;
start_write_param( fs );
cvStartWriteStruct( fs, "power", CV_NODE_SEQ + CV_NODE_FLOW );
for( i = 0; i < CV_DIM(math_pow_values); i++ )
cvWriteReal( fs, 0, math_pow_values[i] );
cvEndWriteStruct(fs);
return code;
}
int CxCore_PowTest::prepare_test_case( int test_case_idx )
{
int code = CxCore_MathTest::prepare_test_case( test_case_idx );
if( code > 0 && ts->get_testing_mode() == CvTS::TIMING_MODE )
{
if( cvRound(power) != power && CV_MAT_DEPTH(test_mat[INPUT][0].type) < CV_32F )
return 0;
}
return code;
}
void CxCore_PowTest::print_timing_params( int test_case_idx, char* ptr, int params_left )
{
sprintf( ptr, "%g,", power );
ptr += strlen(ptr);
params_left--;
CxCore_MathTest::print_timing_params( test_case_idx, ptr, params_left );
}
double CxCore_PowTest::get_success_error_level( int test_case_idx, int i, int j )
{
int type = cvGetElemType( test_array[i][j] );
if( CV_MAT_DEPTH(type) < CV_32F )
return power == cvRound(power) && power >= 0 ? 0 : 1;
else
return CxCore_MathTest::get_success_error_level( test_case_idx, i, j );
}
void CxCore_PowTest::get_minmax_bounds( int /*i*/, int /*j*/, int type, CvScalar* low, CvScalar* high )
{
double l, u = cvTsRandInt(ts->get_rng())%1000 + 1;
if( power > 0 )
{
double mval = cvTsMaxVal(type);
double u1 = pow(mval,1./power)*2;
u = MIN(u,u1);
}
l = power == cvRound(power) ? -u : FLT_EPSILON;
*low = cvScalarAll(l);
*high = cvScalarAll(u);
}
void CxCore_PowTest::run_func()
{
if(!test_nd)
{
if( fabs(power-1./3) <= DBL_EPSILON && CV_MAT_DEPTH(test_mat[INPUT][0].type) == CV_32F )
{
cv::Mat a(&test_mat[INPUT][0]), b(&test_mat[OUTPUT][0]);
a = a.reshape(1);
b = b.reshape(1);
for( int i = 0; i < a.rows; i++ )
{
b.at<float>(i,0) = (float)fabs(cvCbrt(a.at<float>(i,0)));
for( int j = 1; j < a.cols; j++ )
b.at<float>(i,j) = (float)fabs(cv::cubeRoot(a.at<float>(i,j)));
}
}
else
cvPow( test_array[INPUT][0], test_array[OUTPUT][0], power );
}
else
{
cv::MatND a = cv::cvarrToMatND(test_array[INPUT][0]);
cv::MatND b = cv::cvarrToMatND(test_array[OUTPUT][0]);
if(power == 0.5)
cv::sqrt(a, b);
else
cv::pow(a, power, b);
}
}
inline static int ipow( int a, int power )
{
int b = 1;
while( power > 0 )
{
if( power&1 )
b *= a, power--;
else
a *= a, power >>= 1;
}
return b;
}
inline static double ipow( double a, int power )
{
double b = 1.;
while( power > 0 )
{
if( power&1 )
b *= a, power--;
else
a *= a, power >>= 1;
}
return b;
}
void CxCore_PowTest::prepare_to_validation( int /*test_case_idx*/ )
{
CvMat* a = &test_mat[INPUT][0];
CvMat* b = &test_mat[REF_OUTPUT][0];
int depth = CV_MAT_DEPTH(a->type);
int ncols = test_mat[INPUT][0].cols*CV_MAT_CN(a->type);
int ipower = cvRound(power), apower = abs(ipower);
int i, j;
for( i = 0; i < a->rows; i++ )
{
uchar* a_data = a->data.ptr + i*a->step;
uchar* b_data = b->data.ptr + i*b->step;
switch( depth )
{
case CV_8U:
if( ipower < 0 )
for( j = 0; j < ncols; j++ )
{
int val = ((uchar*)a_data)[j];
((uchar*)b_data)[j] = (uchar)(val <= 1 ? val :
val == 2 && ipower == -1 ? 1 : 0);
}
else
for( j = 0; j < ncols; j++ )
{
int val = ((uchar*)a_data)[j];
val = ipow( val, ipower );
((uchar*)b_data)[j] = CV_CAST_8U(val);
}
break;
case CV_8S:
if( ipower < 0 )
for( j = 0; j < ncols; j++ )
{
int val = ((char*)a_data)[j];
((char*)b_data)[j] = (char)((val&~1)==0 ? val :
val ==-1 ? 1-2*(ipower&1) :
val == 2 && ipower == -1 ? 1 : 0);
}
else
for( j = 0; j < ncols; j++ )
{
int val = ((char*)a_data)[j];
val = ipow( val, ipower );
((char*)b_data)[j] = CV_CAST_8S(val);
}
break;
case CV_16U:
if( ipower < 0 )
for( j = 0; j < ncols; j++ )
{
int val = ((ushort*)a_data)[j];
((ushort*)b_data)[j] = (ushort)((val&~1)==0 ? val :
val ==-1 ? 1-2*(ipower&1) :
val == 2 && ipower == -1 ? 1 : 0);
}
else
for( j = 0; j < ncols; j++ )
{
int val = ((ushort*)a_data)[j];
val = ipow( val, ipower );
((ushort*)b_data)[j] = CV_CAST_16U(val);
}
break;
case CV_16S:
if( ipower < 0 )
for( j = 0; j < ncols; j++ )
{
int val = ((short*)a_data)[j];
((short*)b_data)[j] = (short)((val&~1)==0 ? val :
val ==-1 ? 1-2*(ipower&1) :
val == 2 && ipower == -1 ? 1 : 0);
}
else
for( j = 0; j < ncols; j++ )
{
int val = ((short*)a_data)[j];
val = ipow( val, ipower );
((short*)b_data)[j] = CV_CAST_16S(val);
}
break;
case CV_32S:
if( ipower < 0 )
for( j = 0; j < ncols; j++ )
{
int val = ((int*)a_data)[j];
((int*)b_data)[j] = (val&~1)==0 ? val :
val ==-1 ? 1-2*(ipower&1) :
val == 2 && ipower == -1 ? 1 : 0;
}
else
for( j = 0; j < ncols; j++ )
{
int val = ((int*)a_data)[j];
val = ipow( val, ipower );
((int*)b_data)[j] = val;
}
break;
case CV_32F:
if( power != ipower )
for( j = 0; j < ncols; j++ )
{
double val = ((float*)a_data)[j];
val = pow( fabs(val), power );
((float*)b_data)[j] = CV_CAST_32F(val);
}
else
for( j = 0; j < ncols; j++ )
{
double val = ((float*)a_data)[j];
if( ipower < 0 )
val = 1./val;
val = ipow( val, apower );
((float*)b_data)[j] = (float)val;
}
break;
case CV_64F:
if( power != ipower )
for( j = 0; j < ncols; j++ )
{
double val = ((double*)a_data)[j];
val = pow( fabs(val), power );
((double*)b_data)[j] = CV_CAST_64F(val);
}
else
for( j = 0; j < ncols; j++ )
{
double val = ((double*)a_data)[j];
if( ipower < 0 )
val = 1./val;
val = ipow( val, apower );
((double*)b_data)[j] = (double)val;
}
break;
}
}
}
CxCore_PowTest pow_test;
////////// cart2polar /////////////
class CxCore_CartToPolarTest : public CxCore_MathTest
{
public:
CxCore_CartToPolarTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
double get_success_error_level( int test_case_idx, int i, int j );
void run_func();
void prepare_to_validation( int test_case_idx );
int use_degrees;
};
CxCore_CartToPolarTest::CxCore_CartToPolarTest()
: CxCore_MathTest( "math-cart2polar", "cvCartToPolar" )
{
use_degrees = 0;
test_array[INPUT].push(NULL);
test_array[OUTPUT].push(NULL);
test_array[REF_OUTPUT].push(NULL);
}
void CxCore_CartToPolarTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
CxCore_MathTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
use_degrees = cvTsRandInt(rng) & 1;
if( cvTsRandInt(rng) % 4 == 0 ) // check missing magnitude/angle cases
{
int idx = cvTsRandInt(rng) & 1;
sizes[OUTPUT][idx] = sizes[REF_OUTPUT][idx] = cvSize(0,0);
}
}
void CxCore_CartToPolarTest::run_func()
{
if(!test_nd)
{
cvCartToPolar( test_array[INPUT][0], test_array[INPUT][1],
test_array[OUTPUT][0], test_array[OUTPUT][1], use_degrees );
}
else
{
cv::Mat X = cv::cvarrToMat(test_array[INPUT][0]);
cv::Mat Y = cv::cvarrToMat(test_array[INPUT][1]);
cv::Mat mag = test_array[OUTPUT][0] ? cv::cvarrToMat(test_array[OUTPUT][0]) : cv::Mat();
cv::Mat ph = test_array[OUTPUT][1] ? cv::cvarrToMat(test_array[OUTPUT][1]) : cv::Mat();
if(!mag.data)
cv::phase(X, Y, ph, use_degrees != 0);
else if(!ph.data)
cv::magnitude(X, Y, mag);
else
cv::cartToPolar(X, Y, mag, ph, use_degrees != 0);
}
}
double CxCore_CartToPolarTest::get_success_error_level( int test_case_idx, int i, int j )
{
return j == 1 ? 0.5*(use_degrees ? 1 : CV_PI/180.) :
CxCore_MathTest::get_success_error_level( test_case_idx, i, j );
}
void CxCore_CartToPolarTest::prepare_to_validation( int /*test_case_idx*/ )
{
CvMat* x = &test_mat[INPUT][0];
CvMat* y = &test_mat[INPUT][1];
CvMat* mag = test_array[REF_OUTPUT][0] ? &test_mat[REF_OUTPUT][0] : 0;
CvMat* angle = test_array[REF_OUTPUT][1] ? &test_mat[REF_OUTPUT][1] : 0;
double C = use_degrees ? 180./CV_PI : 1.;
int depth = CV_MAT_DEPTH(x->type);
int ncols = x->cols*CV_MAT_CN(x->type);
int i, j;
for( i = 0; i < x->rows; i++ )
{
uchar* x_data = x->data.ptr + i*x->step;
uchar* y_data = y->data.ptr + i*y->step;
uchar* mag_data = mag ? mag->data.ptr + i*mag->step : 0;
uchar* angle_data = angle ? angle->data.ptr + i*angle->step : 0;
if( depth == CV_32F )
{
for( j = 0; j < ncols; j++ )
{
double xval = ((float*)x_data)[j];
double yval = ((float*)y_data)[j];
if( mag_data )
((float*)mag_data)[j] = (float)sqrt(xval*xval + yval*yval);
if( angle_data )
{
double a = atan2( yval, xval );
if( a < 0 )
a += CV_PI*2;
a *= C;
((float*)angle_data)[j] = (float)a;
}
}
}
else
{
assert( depth == CV_64F );
for( j = 0; j < ncols; j++ )
{
double xval = ((double*)x_data)[j];
double yval = ((double*)y_data)[j];
if( mag_data )
((double*)mag_data)[j] = sqrt(xval*xval + yval*yval);
if( angle_data )
{
double a = atan2( yval, xval );
if( a < 0 )
a += CV_PI*2;
a *= C;
((double*)angle_data)[j] = a;
}
}
}
}
if( angle )
{
// hack: increase angle value by 1 (so that alpha becomes 1+alpha)
// to hide large relative errors in case of very small angles
cvTsAdd( &test_mat[OUTPUT][1], cvScalarAll(1.), 0, cvScalarAll(0.),
cvScalarAll(1.), &test_mat[OUTPUT][1], 0 );
cvTsAdd( &test_mat[REF_OUTPUT][1], cvScalarAll(1.), 0, cvScalarAll(0.),
cvScalarAll(1.), &test_mat[REF_OUTPUT][1], 0 );
}
}
CxCore_CartToPolarTest cart2polar_test;
////////// polar2cart /////////////
class CxCore_PolarToCartTest : public CxCore_MathTest
{
public:
CxCore_PolarToCartTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
double get_success_error_level( int test_case_idx, int i, int j );
void run_func();
void prepare_to_validation( int test_case_idx );
int use_degrees;
};
CxCore_PolarToCartTest::CxCore_PolarToCartTest()
: CxCore_MathTest( "math-polar2cart", "cvPolarToCart" )
{
use_degrees = 0;
test_array[INPUT].push(NULL);
test_array[OUTPUT].push(NULL);
test_array[REF_OUTPUT].push(NULL);
}
void CxCore_PolarToCartTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
CxCore_MathTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
use_degrees = cvTsRandInt(rng) & 1;
if( cvTsRandInt(rng) % 4 == 0 ) // check missing magnitude case
sizes[INPUT][1] = cvSize(0,0);
if( cvTsRandInt(rng) % 4 == 0 ) // check missing x/y cases
{
int idx = cvTsRandInt(rng) & 1;
sizes[OUTPUT][idx] = sizes[REF_OUTPUT][idx] = cvSize(0,0);
}
}
void CxCore_PolarToCartTest::run_func()
{
if(!test_nd)
{
cvPolarToCart( test_array[INPUT][1], test_array[INPUT][0],
test_array[OUTPUT][0], test_array[OUTPUT][1], use_degrees );
}
else
{
cv::Mat X = test_array[OUTPUT][0] ? cv::cvarrToMat(test_array[OUTPUT][0]) : cv::Mat();
cv::Mat Y = test_array[OUTPUT][1] ? cv::cvarrToMat(test_array[OUTPUT][1]) : cv::Mat();
cv::Mat mag = test_array[INPUT][1] ? cv::cvarrToMat(test_array[INPUT][1]) : cv::Mat();
cv::Mat ph = test_array[INPUT][0] ? cv::cvarrToMat(test_array[INPUT][0]) : cv::Mat();
cv::polarToCart(mag, ph, X, Y, use_degrees != 0);
}
}
double CxCore_PolarToCartTest::get_success_error_level( int /*test_case_idx*/, int /*i*/, int /*j*/ )
{
return FLT_EPSILON*100;
}
void CxCore_PolarToCartTest::prepare_to_validation( int /*test_case_idx*/ )
{
CvMat* x = test_array[REF_OUTPUT][0] ? &test_mat[REF_OUTPUT][0] : 0;
CvMat* y = test_array[REF_OUTPUT][1] ? &test_mat[REF_OUTPUT][1] : 0;
CvMat* angle = &test_mat[INPUT][0];
CvMat* mag = test_array[INPUT][1] ? &test_mat[INPUT][1] : 0;
double C = use_degrees ? CV_PI/180. : 1.;
int depth = CV_MAT_DEPTH(angle->type);
int ncols = angle->cols*CV_MAT_CN(angle->type);
int i, j;
for( i = 0; i < angle->rows; i++ )
{
uchar* x_data = x ? x->data.ptr + i*x->step : 0;
uchar* y_data = y ? y->data.ptr + i*y->step : 0;
uchar* mag_data = mag ? mag->data.ptr + i*mag->step : 0;
uchar* angle_data = angle->data.ptr + i*angle->step;
if( depth == CV_32F )
{
for( j = 0; j < ncols; j++ )
{
double a = ((float*)angle_data)[j]*C;
double m = mag_data ? ((float*)mag_data)[j] : 1.;
if( x_data )
((float*)x_data)[j] = (float)(m*cos(a));
if( y_data )
((float*)y_data)[j] = (float)(m*sin(a));
}
}
else
{
assert( depth == CV_64F );
for( j = 0; j < ncols; j++ )
{
double a = ((double*)angle_data)[j]*C;
double m = mag_data ? ((double*)mag_data)[j] : 1.;
if( x_data )
((double*)x_data)[j] = m*cos(a);
if( y_data )
((double*)y_data)[j] = m*sin(a);
}
}
}
}
CxCore_PolarToCartTest polar2cart_test;
///////////////////////////////////////// matrix tests ////////////////////////////////////////////
static const int matrix_all_depths[] = { CV_8U, CV_16U, CV_16S, CV_32S, CV_32F, CV_64F, -1 };
class CxCore_MatrixTestImpl : public CvArrTest
{
public:
CxCore_MatrixTestImpl( const char* test_name, const char* test_funcs, int in_count, int out_count,
bool allow_int, bool scalar_output, int max_cn );
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images );
double get_success_error_level( int test_case_idx, int i, int j );
bool allow_int;
bool scalar_output;
int max_cn;
};
CxCore_MatrixTestImpl::CxCore_MatrixTestImpl( const char* test_name, const char* test_funcs,
int in_count, int out_count,
bool _allow_int, bool _scalar_output, int _max_cn )
: CvArrTest( test_name, test_funcs, "" ),
allow_int(_allow_int), scalar_output(_scalar_output), max_cn(_max_cn)
{
int i;
for( i = 0; i < in_count; i++ )
test_array[INPUT].push(NULL);
for( i = 0; i < out_count; i++ )
{
test_array[OUTPUT].push(NULL);
test_array[REF_OUTPUT].push(NULL);
}
element_wise_relative_error = false;
default_timing_param_names = math_param_names;
size_list = (CvSize*)matrix_sizes;
whole_size_list = 0;
depth_list = (int*)math_depths;
cn_list = 0;
}
void CxCore_MatrixTestImpl::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
int depth = cvTsRandInt(rng) % (allow_int ? CV_64F+1 : 2);
int cn = cvTsRandInt(rng) % max_cn + 1;
int i, j;
if( allow_int )
depth += depth == CV_8S;
else
depth += CV_32F;
CvArrTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
for( i = 0; i < max_arr; i++ )
{
int count = test_array[i].size();
int flag = (i == OUTPUT || i == REF_OUTPUT) && scalar_output;
int type = !flag ? CV_MAKETYPE(depth, cn) : CV_64FC1;
for( j = 0; j < count; j++ )
{
types[i][j] = type;
if( flag )
sizes[i][j] = cvSize( 4, 1 );
}
}
}
void CxCore_MatrixTestImpl::get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types, CvSize** whole_sizes, bool* are_images )
{
CvArrTest::get_timing_test_array_types_and_sizes( test_case_idx,
sizes, types, whole_sizes, are_images );
if( scalar_output )
{
types[OUTPUT][0] = types[REF_OUTPUT][0] = CV_64FC1;
sizes[OUTPUT][0] = sizes[REF_OUTPUT][0] = cvSize( 4, 1 );
whole_sizes[OUTPUT][0] = whole_sizes[REF_OUTPUT][0] = cvSize( 4, 1 );
}
}
double CxCore_MatrixTestImpl::get_success_error_level( int test_case_idx, int i, int j )
{
int input_depth = CV_MAT_DEPTH(cvGetElemType( test_array[INPUT][0] ));
double input_precision = input_depth < CV_32F ? 0 : input_depth == CV_32F ?
5e-5 : 1e-10;
double output_precision = CvArrTest::get_success_error_level( test_case_idx, i, j );
return MAX(input_precision, output_precision);
}
CxCore_MatrixTestImpl matrix_test( "matrix", "", 0, 0, false, false, 0 );
class CxCore_MatrixTest : public CxCore_MatrixTestImpl
{
public:
CxCore_MatrixTest( const char* test_name, const char* test_funcs, int in_count, int out_count,
bool allow_int, bool scalar_output, int max_cn );
};
CxCore_MatrixTest::CxCore_MatrixTest( const char* test_name, const char* test_funcs,
int in_count, int out_count, bool _allow_int,
bool _scalar_output, int _max_cn )
: CxCore_MatrixTestImpl( test_name, test_funcs, in_count, out_count,
_allow_int, _scalar_output, _max_cn )
{
size_list = 0;
depth_list = 0;
}
///////////////// Trace /////////////////////
class CxCore_TraceTest : public CxCore_MatrixTest
{
public:
CxCore_TraceTest();
protected:
void run_func();
void prepare_to_validation( int test_case_idx );
};
CxCore_TraceTest::CxCore_TraceTest() :
CxCore_MatrixTest( "matrix-trace", "cvTrace", 1, 1, true, true, 4 )
{
}
void CxCore_TraceTest::run_func()
{
*((CvScalar*)(test_mat[OUTPUT][0].data.db)) = cvTrace(test_array[INPUT][0]);
}
void CxCore_TraceTest::prepare_to_validation( int )
{
CvMat* mat = &test_mat[INPUT][0];
int i, j, count = MIN( mat->rows, mat->cols );
CvScalar trace = {{0,0,0,0}};
for( i = 0; i < count; i++ )
{
CvScalar el = cvGet2D( mat, i, i );
for( j = 0; j < 4; j++ )
trace.val[j] += el.val[j];
}
*((CvScalar*)(test_mat[REF_OUTPUT][0].data.db)) = trace;
}
CxCore_TraceTest trace_test;
///////// dotproduct //////////
class CxCore_DotProductTest : public CxCore_MatrixTest
{
public:
CxCore_DotProductTest();
protected:
void run_func();
void prepare_to_validation( int test_case_idx );
};
CxCore_DotProductTest::CxCore_DotProductTest() :
CxCore_MatrixTest( "matrix-dotproduct", "cvDotProduct", 2, 1, true, true, 4 )
{
depth_list = matrix_all_depths;
}
void CxCore_DotProductTest::run_func()
{
*((CvScalar*)(test_mat[OUTPUT][0].data.ptr)) =
cvRealScalar(cvDotProduct( test_array[INPUT][0], test_array[INPUT][1] ));
}
void CxCore_DotProductTest::prepare_to_validation( int )
{
*((CvScalar*)(test_mat[REF_OUTPUT][0].data.ptr)) =
cvRealScalar(cvTsCrossCorr( &test_mat[INPUT][0], &test_mat[INPUT][1] ));
}
CxCore_DotProductTest dotproduct_test;
///////// crossproduct //////////
static const CvSize cross_product_sizes[] = {{3,1}, {-1,-1}};
class CxCore_CrossProductTest : public CxCore_MatrixTest
{
public:
CxCore_CrossProductTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void run_func();
void prepare_to_validation( int test_case_idx );
};
CxCore_CrossProductTest::CxCore_CrossProductTest() :
CxCore_MatrixTest( "matrix-crossproduct", "cvCrossProduct", 2, 1, false, false, 1 )
{
size_list = cross_product_sizes;
}
void CxCore_CrossProductTest::get_test_array_types_and_sizes( int /*test_case_idx*/, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
int depth = cvTsRandInt(rng) % 2 + CV_32F;
int cn = cvTsRandInt(rng) & 1 ? 3 : 1, type = CV_MAKETYPE(depth, cn);
CvSize sz;
types[INPUT][0] = types[INPUT][1] = types[OUTPUT][0] = types[REF_OUTPUT][0] = type;
if( cn == 3 )
sz = cvSize(1,1);
else if( cvTsRandInt(rng) & 1 )
sz = cvSize(3,1);
else
sz = cvSize(1,3);
sizes[INPUT][0] = sizes[INPUT][1] = sizes[OUTPUT][0] = sizes[REF_OUTPUT][0] = sz;
}
void CxCore_CrossProductTest::run_func()
{
cvCrossProduct( test_array[INPUT][0], test_array[INPUT][1], test_array[OUTPUT][0] );
}
void CxCore_CrossProductTest::prepare_to_validation( int )
{
CvScalar a = {{0,0,0,0}}, b = {{0,0,0,0}}, c = {{0,0,0,0}};
if( test_mat[INPUT][0].rows > 1 )
{
a.val[0] = cvGetReal2D( &test_mat[INPUT][0], 0, 0 );
a.val[1] = cvGetReal2D( &test_mat[INPUT][0], 1, 0 );
a.val[2] = cvGetReal2D( &test_mat[INPUT][0], 2, 0 );
b.val[0] = cvGetReal2D( &test_mat[INPUT][1], 0, 0 );
b.val[1] = cvGetReal2D( &test_mat[INPUT][1], 1, 0 );
b.val[2] = cvGetReal2D( &test_mat[INPUT][1], 2, 0 );
}
else if( test_mat[INPUT][0].cols > 1 )
{
a.val[0] = cvGetReal1D( &test_mat[INPUT][0], 0 );
a.val[1] = cvGetReal1D( &test_mat[INPUT][0], 1 );
a.val[2] = cvGetReal1D( &test_mat[INPUT][0], 2 );
b.val[0] = cvGetReal1D( &test_mat[INPUT][1], 0 );
b.val[1] = cvGetReal1D( &test_mat[INPUT][1], 1 );
b.val[2] = cvGetReal1D( &test_mat[INPUT][1], 2 );
}
else
{
a = cvGet1D( &test_mat[INPUT][0], 0 );
b = cvGet1D( &test_mat[INPUT][1], 0 );
}
c.val[2] = a.val[0]*b.val[1] - a.val[1]*b.val[0];
c.val[1] = -a.val[0]*b.val[2] + a.val[2]*b.val[0];
c.val[0] = a.val[1]*b.val[2] - a.val[2]*b.val[1];
if( test_mat[REF_OUTPUT][0].rows > 1 )
{
cvSetReal2D( &test_mat[REF_OUTPUT][0], 0, 0, c.val[0] );
cvSetReal2D( &test_mat[REF_OUTPUT][0], 1, 0, c.val[1] );
cvSetReal2D( &test_mat[REF_OUTPUT][0], 2, 0, c.val[2] );
}
else if( test_mat[REF_OUTPUT][0].cols > 1 )
{
cvSetReal1D( &test_mat[REF_OUTPUT][0], 0, c.val[0] );
cvSetReal1D( &test_mat[REF_OUTPUT][0], 1, c.val[1] );
cvSetReal1D( &test_mat[REF_OUTPUT][0], 2, c.val[2] );
}
else
{
cvSet1D( &test_mat[REF_OUTPUT][0], 0, c );
}
}
CxCore_CrossProductTest crossproduct_test;
///////////////// scaleadd /////////////////////
class CxCore_ScaleAddTest : public CxCore_MatrixTest
{
public:
CxCore_ScaleAddTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images );
int prepare_test_case( int test_case_idx );
void run_func();
void prepare_to_validation( int test_case_idx );
CvScalar alpha;
bool test_nd;
};
CxCore_ScaleAddTest::CxCore_ScaleAddTest() :
CxCore_MatrixTest( "matrix-scaleadd", "cvScaleAdd", 3, 1, false, false, 4 )
{
alpha = cvScalarAll(0);
test_nd = false;
}
void CxCore_ScaleAddTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
sizes[INPUT][2] = cvSize(1,1);
types[INPUT][2] &= CV_MAT_DEPTH_MASK;
test_nd = cvTsRandInt(ts->get_rng()) % 2 != 0;
}
void CxCore_ScaleAddTest::get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images )
{
CxCore_MatrixTest::get_timing_test_array_types_and_sizes( test_case_idx, sizes, types,
whole_sizes, are_images );
sizes[INPUT][2] = cvSize(1,1);
types[INPUT][2] &= CV_MAT_DEPTH_MASK;
}
int CxCore_ScaleAddTest::prepare_test_case( int test_case_idx )
{
int code = CxCore_MatrixTest::prepare_test_case( test_case_idx );
if( code > 0 )
alpha = cvGet1D( &test_mat[INPUT][2], 0 );
if( test_nd )
alpha.val[1] = 0;
return code;
}
void CxCore_ScaleAddTest::run_func()
{
if(!test_nd)
cvScaleAdd( test_array[INPUT][0], alpha, test_array[INPUT][1], test_array[OUTPUT][0] );
else
{
cv::MatND c = cv::cvarrToMatND(test_array[OUTPUT][0]);
cv::scaleAdd( cv::cvarrToMatND(test_array[INPUT][0]), alpha.val[0],
cv::cvarrToMatND(test_array[INPUT][1]), c);
}
}
void CxCore_ScaleAddTest::prepare_to_validation( int )
{
cvTsAdd( &test_mat[INPUT][0], cvScalarAll(alpha.val[0]),
&test_mat[INPUT][1], cvScalarAll(1.),
cvScalarAll(0.), &test_mat[REF_OUTPUT][0], 0 );
}
CxCore_ScaleAddTest scaleadd_test;
///////////////// gemm /////////////////////
static const char* matrix_gemm_param_names[] = { "size", "add_c", "mul_type", "depth", 0 };
static const char* matrix_gemm_mul_types[] = { "AB", "AtB", "ABt", "AtBt", 0 };
static const int matrix_gemm_add_c_flags[] = { 0, 1 };
class CxCore_GEMMTest : public CxCore_MatrixTest
{
public:
CxCore_GEMMTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high );
void get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images );
int write_default_params( CvFileStorage* fs );
void print_timing_params( int test_case_idx, char* ptr, int params_left );
int prepare_test_case( int test_case_idx );
void run_func();
void prepare_to_validation( int test_case_idx );
int tabc_flag;
double alpha, beta;
};
CxCore_GEMMTest::CxCore_GEMMTest() :
CxCore_MatrixTest( "matrix-gemm", "cvGEMM", 5, 1, false, false, 2 )
{
test_case_count = 100;
max_log_array_size = 10;
default_timing_param_names = matrix_gemm_param_names;
alpha = beta = 0;
}
void CxCore_GEMMTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
CvSize sizeA;
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
sizeA = sizes[INPUT][0];
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
sizes[INPUT][0] = sizeA;
sizes[INPUT][2] = sizes[INPUT][3] = cvSize(1,1);
types[INPUT][2] = types[INPUT][3] &= ~CV_MAT_CN_MASK;
tabc_flag = cvTsRandInt(rng) & 7;
switch( tabc_flag & (CV_GEMM_A_T|CV_GEMM_B_T) )
{
case 0:
sizes[INPUT][1].height = sizes[INPUT][0].width;
sizes[OUTPUT][0].height = sizes[INPUT][0].height;
sizes[OUTPUT][0].width = sizes[INPUT][1].width;
break;
case CV_GEMM_B_T:
sizes[INPUT][1].width = sizes[INPUT][0].width;
sizes[OUTPUT][0].height = sizes[INPUT][0].height;
sizes[OUTPUT][0].width = sizes[INPUT][1].height;
break;
case CV_GEMM_A_T:
sizes[INPUT][1].height = sizes[INPUT][0].height;
sizes[OUTPUT][0].height = sizes[INPUT][0].width;
sizes[OUTPUT][0].width = sizes[INPUT][1].width;
break;
case CV_GEMM_A_T | CV_GEMM_B_T:
sizes[INPUT][1].width = sizes[INPUT][0].height;
sizes[OUTPUT][0].height = sizes[INPUT][0].width;
sizes[OUTPUT][0].width = sizes[INPUT][1].height;
break;
}
sizes[REF_OUTPUT][0] = sizes[OUTPUT][0];
if( cvTsRandInt(rng) & 1 )
sizes[INPUT][4] = cvSize(0,0);
else if( !(tabc_flag & CV_GEMM_C_T) )
sizes[INPUT][4] = sizes[OUTPUT][0];
else
{
sizes[INPUT][4].width = sizes[OUTPUT][0].height;
sizes[INPUT][4].height = sizes[OUTPUT][0].width;
}
}
void CxCore_GEMMTest::get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images )
{
CxCore_MatrixTest::get_timing_test_array_types_and_sizes( test_case_idx,
sizes, types, whole_sizes, are_images );
const char* mul_type = cvReadString( find_timing_param("mul_type"), "AB" );
if( strcmp( mul_type, "AtB" ) == 0 )
tabc_flag = CV_GEMM_A_T;
else if( strcmp( mul_type, "ABt" ) == 0 )
tabc_flag = CV_GEMM_B_T;
else if( strcmp( mul_type, "AtBt" ) == 0 )
tabc_flag = CV_GEMM_A_T + CV_GEMM_B_T;
else
tabc_flag = 0;
if( cvReadInt( find_timing_param( "add_c" ), 0 ) == 0 )
sizes[INPUT][4] = cvSize(0,0);
}
int CxCore_GEMMTest::write_default_params( CvFileStorage* fs )
{
int code = CxCore_MatrixTest::write_default_params(fs);
if( code < 0 || ts->get_testing_mode() != CvTS::TIMING_MODE )
return code;
write_string_list( fs, "mul_type", matrix_gemm_mul_types );
write_int_list( fs, "add_c", matrix_gemm_add_c_flags, CV_DIM(matrix_gemm_add_c_flags) );
return code;
}
void CxCore_GEMMTest::print_timing_params( int test_case_idx, char* ptr, int params_left )
{
sprintf( ptr, "%s%s,%s,",
tabc_flag & CV_GEMM_A_T ? "At" : "A",
tabc_flag & CV_GEMM_B_T ? "Bt" : "B",
test_array[INPUT][4] ? "plusC" : "" );
ptr += strlen(ptr);
params_left -= 2;
CxCore_MatrixTest::print_timing_params( test_case_idx, ptr, params_left );
}
int CxCore_GEMMTest::prepare_test_case( int test_case_idx )
{
int code = CxCore_MatrixTest::prepare_test_case( test_case_idx );
if( code > 0 )
{
alpha = cvmGet( &test_mat[INPUT][2], 0, 0 );
beta = cvmGet( &test_mat[INPUT][3], 0, 0 );
}
return code;
}
void CxCore_GEMMTest::get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high )
{
*low = cvScalarAll(-10.);
*high = cvScalarAll(10.);
}
void CxCore_GEMMTest::run_func()
{
cvGEMM( test_array[INPUT][0], test_array[INPUT][1], alpha,
test_array[INPUT][4], beta, test_array[OUTPUT][0], tabc_flag );
}
void CxCore_GEMMTest::prepare_to_validation( int )
{
cvTsGEMM( &test_mat[INPUT][0], &test_mat[INPUT][1], alpha,
test_array[INPUT][4] ? &test_mat[INPUT][4] : 0,
beta, &test_mat[REF_OUTPUT][0], tabc_flag );
}
CxCore_GEMMTest gemm_test;
///////////////// multransposed /////////////////////
static const char* matrix_multrans_param_names[] = { "size", "use_delta", "mul_type", "depth", 0 };
static const int matrix_multrans_use_delta_flags[] = { 0, 1 };
static const char* matrix_multrans_mul_types[] = { "AAt", "AtA", 0 };
class CxCore_MulTransposedTest : public CxCore_MatrixTest
{
public:
CxCore_MulTransposedTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images );
int write_default_params( CvFileStorage* fs );
void print_timing_params( int test_case_idx, char* ptr, int params_left );
void get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high );
void run_func();
void prepare_to_validation( int test_case_idx );
int order;
};
CxCore_MulTransposedTest::CxCore_MulTransposedTest() :
CxCore_MatrixTest( "matrix-multransposed", "cvMulTransposed, cvRepeat", 2, 1, false, false, 1 )
{
test_case_count = 100;
order = 0;
test_array[TEMP].push(NULL);
default_timing_param_names = matrix_multrans_param_names;
}
void CxCore_MulTransposedTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
int bits = cvTsRandInt(rng);
int src_type = cvTsRandInt(rng) % 5;
int dst_type = cvTsRandInt(rng) % 2;
src_type = src_type == 0 ? CV_8U : src_type == 1 ? CV_16U : src_type == 2 ? CV_16S :
src_type == 3 ? CV_32F : CV_64F;
dst_type = dst_type == 0 ? CV_32F : CV_64F;
dst_type = MAX( dst_type, src_type );
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
if( bits & 1 )
sizes[INPUT][1] = cvSize(0,0);
else
{
sizes[INPUT][1] = sizes[INPUT][0];
if( bits & 2 )
sizes[INPUT][1].height = 1;
if( bits & 4 )
sizes[INPUT][1].width = 1;
}
sizes[TEMP][0] = sizes[INPUT][0];
types[INPUT][0] = src_type;
types[OUTPUT][0] = types[REF_OUTPUT][0] = types[INPUT][1] = types[TEMP][0] = dst_type;
order = (bits & 8) != 0;
sizes[OUTPUT][0].width = sizes[OUTPUT][0].height = order == 0 ?
sizes[INPUT][0].height : sizes[INPUT][0].width;
sizes[REF_OUTPUT][0] = sizes[OUTPUT][0];
}
void CxCore_MulTransposedTest::get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images )
{
CxCore_MatrixTest::get_timing_test_array_types_and_sizes( test_case_idx,
sizes, types, whole_sizes, are_images );
const char* mul_type = cvReadString( find_timing_param("mul_type"), "AAt" );
order = strcmp( mul_type, "AtA" ) == 0;
if( cvReadInt( find_timing_param( "use_delta" ), 0 ) == 0 )
sizes[INPUT][1] = cvSize(0,0);
}
int CxCore_MulTransposedTest::write_default_params( CvFileStorage* fs )
{
int code = CxCore_MatrixTest::write_default_params(fs);
if( code < 0 || ts->get_testing_mode() != CvTS::TIMING_MODE )
return code;
write_string_list( fs, "mul_type", matrix_multrans_mul_types );
write_int_list( fs, "use_delta", matrix_multrans_use_delta_flags,
CV_DIM(matrix_multrans_use_delta_flags) );
return code;
}
void CxCore_MulTransposedTest::print_timing_params( int test_case_idx, char* ptr, int params_left )
{
sprintf( ptr, "%s,%s,", order == 0 ? "AAt" : "AtA", test_array[INPUT][1] ? "delta" : "" );
ptr += strlen(ptr);
params_left -= 2;
CxCore_MatrixTest::print_timing_params( test_case_idx, ptr, params_left );
}
void CxCore_MulTransposedTest::get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high )
{
*low = cvScalarAll(-10.);
*high = cvScalarAll(10.);
}
void CxCore_MulTransposedTest::run_func()
{
cvMulTransposed( test_array[INPUT][0], test_array[OUTPUT][0],
order, test_array[INPUT][1] );
}
void CxCore_MulTransposedTest::prepare_to_validation( int )
{
CvMat* delta = test_array[INPUT][1] ? &test_mat[INPUT][1] : 0;
if( delta )
{
if( test_mat[INPUT][1].rows < test_mat[INPUT][0].rows ||
test_mat[INPUT][1].cols < test_mat[INPUT][0].cols )
{
cvRepeat( delta, &test_mat[TEMP][0] );
delta = &test_mat[TEMP][0];
}
cvTsAdd( &test_mat[INPUT][0], cvScalarAll(1.), delta, cvScalarAll(-1.),
cvScalarAll(0.), &test_mat[TEMP][0], 0 );
}
else
cvTsConvert( &test_mat[INPUT][0], &test_mat[TEMP][0] );
delta = &test_mat[TEMP][0];
cvTsGEMM( delta, delta, 1., 0, 0, &test_mat[REF_OUTPUT][0], order == 0 ? CV_GEMM_B_T : CV_GEMM_A_T );
}
CxCore_MulTransposedTest multransposed_test;
///////////////// Transform /////////////////////
static const CvSize matrix_transform_sizes[] = {{10,10}, {100,100}, {720,480}, {-1,-1}};
static const CvSize matrix_transform_whole_sizes[] = {{10,10}, {720,480}, {720,480}, {-1,-1}};
static const int matrix_transform_channels[] = { 2, 3, 4, -1 };
static const char* matrix_transform_param_names[] = { "size", "channels", "depth", 0 };
class CxCore_TransformTest : public CxCore_MatrixTest
{
public:
CxCore_TransformTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
double get_success_error_level( int test_case_idx, int i, int j );
void get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images );
int prepare_test_case( int test_case_idx );
void print_timing_params( int test_case_idx, char* ptr, int params_left );
void run_func();
void prepare_to_validation( int test_case_idx );
double scale;
bool diagMtx;
};
CxCore_TransformTest::CxCore_TransformTest() :
CxCore_MatrixTest( "matrix-transform", "cvTransform", 3, 1, true, false, 4 )
{
default_timing_param_names = matrix_transform_param_names;
cn_list = matrix_transform_channels;
depth_list = matrix_all_depths;
size_list = matrix_transform_sizes;
whole_size_list = matrix_transform_whole_sizes;
}
void CxCore_TransformTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
int bits = cvTsRandInt(rng);
int depth, dst_cn, mat_cols, mattype;
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
mat_cols = CV_MAT_CN(types[INPUT][0]);
depth = CV_MAT_DEPTH(types[INPUT][0]);
dst_cn = cvTsRandInt(rng) % 4 + 1;
types[OUTPUT][0] = types[REF_OUTPUT][0] = CV_MAKETYPE(depth, dst_cn);
mattype = depth < CV_32S ? CV_32F : depth == CV_64F ? CV_64F : bits & 1 ? CV_32F : CV_64F;
types[INPUT][1] = mattype;
types[INPUT][2] = CV_MAKETYPE(mattype, dst_cn);
scale = 1./((cvTsRandInt(rng)%4)*50+1);
if( bits & 2 )
{
sizes[INPUT][2] = cvSize(0,0);
mat_cols += (bits & 4) != 0;
}
else if( bits & 4 )
sizes[INPUT][2] = cvSize(1,1);
else
{
if( bits & 8 )
sizes[INPUT][2] = cvSize(dst_cn,1);
else
sizes[INPUT][2] = cvSize(1,dst_cn);
types[INPUT][2] &= ~CV_MAT_CN_MASK;
}
diagMtx = (bits & 16) != 0;
sizes[INPUT][1] = cvSize(mat_cols,dst_cn);
}
void CxCore_TransformTest::get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types, CvSize** whole_sizes, bool* are_images )
{
CxCore_MatrixTest::get_timing_test_array_types_and_sizes( test_case_idx,
sizes, types, whole_sizes, are_images );
int cn = CV_MAT_CN(types[INPUT][0]);
sizes[INPUT][1] = cvSize(cn + (cn < 4), cn);
sizes[INPUT][2] = cvSize(0,0);
types[INPUT][1] = types[INPUT][2] = CV_64FC1;
scale = 1./1000;
}
int CxCore_TransformTest::prepare_test_case( int test_case_idx )
{
int code = CxCore_MatrixTest::prepare_test_case( test_case_idx );
if( code > 0 )
{
cvTsAdd(&test_mat[INPUT][1], cvScalarAll(scale), &test_mat[INPUT][1],
cvScalarAll(0), cvScalarAll(0), &test_mat[INPUT][1], 0 );
if(diagMtx)
{
CvMat* w = cvCloneMat(&test_mat[INPUT][1]);
cvSetIdentity(w, cvScalarAll(1));
cvMul(w, &test_mat[INPUT][1], &test_mat[INPUT][1]);
cvReleaseMat(&w);
}
}
return code;
}
void CxCore_TransformTest::print_timing_params( int test_case_idx, char* ptr, int params_left )
{
CvSize size = cvGetMatSize(&test_mat[INPUT][1]);
sprintf( ptr, "matrix=%dx%d,", size.height, size.width );
ptr += strlen(ptr);
params_left--;
CxCore_MatrixTest::print_timing_params( test_case_idx, ptr, params_left );
}
double CxCore_TransformTest::get_success_error_level( int test_case_idx, int i, int j )
{
int depth = CV_MAT_DEPTH(test_mat[INPUT][0].type);
return depth <= CV_8S ? 1 : depth <= CV_32S ? 9 :
CxCore_MatrixTest::get_success_error_level( test_case_idx, i, j );
}
void CxCore_TransformTest::run_func()
{
cvTransform( test_array[INPUT][0], test_array[OUTPUT][0], &test_mat[INPUT][1],
test_array[INPUT][2] ? &test_mat[INPUT][2] : 0);
}
void CxCore_TransformTest::prepare_to_validation( int )
{
CvMat* transmat = &test_mat[INPUT][1];
CvMat* shift = test_array[INPUT][2] ? &test_mat[INPUT][2] : 0;
cvTsTransform( &test_mat[INPUT][0], &test_mat[REF_OUTPUT][0], transmat, shift );
}
CxCore_TransformTest transform_test;
///////////////// PerspectiveTransform /////////////////////
static const int matrix_perspective_transform_channels[] = { 2, 3, -1 };
class CxCore_PerspectiveTransformTest : public CxCore_MatrixTest
{
public:
CxCore_PerspectiveTransformTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
double get_success_error_level( int test_case_idx, int i, int j );
void get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images );
void run_func();
void prepare_to_validation( int test_case_idx );
};
CxCore_PerspectiveTransformTest::CxCore_PerspectiveTransformTest() :
CxCore_MatrixTest( "matrix-perspective", "cvPerspectiveTransform", 2, 1, false, false, 2 )
{
default_timing_param_names = matrix_transform_param_names;
cn_list = matrix_perspective_transform_channels;
size_list = matrix_transform_sizes;
whole_size_list = matrix_transform_whole_sizes;
}
void CxCore_PerspectiveTransformTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
int bits = cvTsRandInt(rng);
int depth, cn, mattype;
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
cn = CV_MAT_CN(types[INPUT][0]) + 1;
depth = CV_MAT_DEPTH(types[INPUT][0]);
types[INPUT][0] = types[OUTPUT][0] = types[REF_OUTPUT][0] = CV_MAKETYPE(depth, cn);
mattype = depth == CV_64F ? CV_64F : bits & 1 ? CV_32F : CV_64F;
types[INPUT][1] = mattype;
sizes[INPUT][1] = cvSize(cn + 1, cn + 1);
}
double CxCore_PerspectiveTransformTest::get_success_error_level( int test_case_idx, int i, int j )
{
int depth = CV_MAT_DEPTH(test_mat[INPUT][0].type);
return depth == CV_32F ? 1e-4 : depth == CV_64F ? 1e-8 :
CxCore_MatrixTest::get_success_error_level(test_case_idx, i, j);
}
void CxCore_PerspectiveTransformTest::get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types, CvSize** whole_sizes, bool* are_images )
{
CxCore_MatrixTest::get_timing_test_array_types_and_sizes( test_case_idx,
sizes, types, whole_sizes, are_images );
int cn = CV_MAT_CN(types[INPUT][0]);
sizes[INPUT][1] = cvSize(cn + 1, cn + 1);
types[INPUT][1] = CV_64FC1;
}
void CxCore_PerspectiveTransformTest::run_func()
{
cvPerspectiveTransform( test_array[INPUT][0], test_array[OUTPUT][0], &test_mat[INPUT][1] );
}
static void cvTsPerspectiveTransform( const CvArr* _src, CvArr* _dst, const CvMat* transmat )
{
int i, j, cols;
int cn, depth, mat_depth;
CvMat astub, bstub, *a, *b;
double mat[16], *buf;
a = cvGetMat( _src, &astub, 0, 0 );
b = cvGetMat( _dst, &bstub, 0, 0 );
cn = CV_MAT_CN(a->type);
depth = CV_MAT_DEPTH(a->type);
mat_depth = CV_MAT_DEPTH(transmat->type);
cols = transmat->cols;
// prepare cn x (cn + 1) transform matrix
if( mat_depth == CV_32F )
{
for( i = 0; i < transmat->rows; i++ )
for( j = 0; j < cols; j++ )
mat[i*cols + j] = ((float*)(transmat->data.ptr + transmat->step*i))[j];
}
else
{
assert( mat_depth == CV_64F );
for( i = 0; i < transmat->rows; i++ )
for( j = 0; j < cols; j++ )
mat[i*cols + j] = ((double*)(transmat->data.ptr + transmat->step*i))[j];
}
// transform data
cols = a->cols * cn;
buf = (double*)cvStackAlloc( cols * sizeof(double) );
for( i = 0; i < a->rows; i++ )
{
uchar* src = a->data.ptr + i*a->step;
uchar* dst = b->data.ptr + i*b->step;
switch( depth )
{
case CV_32F:
for( j = 0; j < cols; j++ )
buf[j] = ((float*)src)[j];
break;
case CV_64F:
for( j = 0; j < cols; j++ )
buf[j] = ((double*)src)[j];
break;
default:
assert(0);
}
switch( cn )
{
case 2:
for( j = 0; j < cols; j += 2 )
{
double t0 = buf[j]*mat[0] + buf[j+1]*mat[1] + mat[2];
double t1 = buf[j]*mat[3] + buf[j+1]*mat[4] + mat[5];
double w = buf[j]*mat[6] + buf[j+1]*mat[7] + mat[8];
w = w ? 1./w : 0;
buf[j] = t0*w;
buf[j+1] = t1*w;
}
break;
case 3:
for( j = 0; j < cols; j += 3 )
{
double t0 = buf[j]*mat[0] + buf[j+1]*mat[1] + buf[j+2]*mat[2] + mat[3];
double t1 = buf[j]*mat[4] + buf[j+1]*mat[5] + buf[j+2]*mat[6] + mat[7];
double t2 = buf[j]*mat[8] + buf[j+1]*mat[9] + buf[j+2]*mat[10] + mat[11];
double w = buf[j]*mat[12] + buf[j+1]*mat[13] + buf[j+2]*mat[14] + mat[15];
w = w ? 1./w : 0;
buf[j] = t0*w;
buf[j+1] = t1*w;
buf[j+2] = t2*w;
}
break;
default:
assert(0);
}
switch( depth )
{
case CV_32F:
for( j = 0; j < cols; j++ )
((float*)dst)[j] = (float)buf[j];
break;
case CV_64F:
for( j = 0; j < cols; j++ )
((double*)dst)[j] = buf[j];
break;
default:
assert(0);
}
}
}
void CxCore_PerspectiveTransformTest::prepare_to_validation( int )
{
CvMat* transmat = &test_mat[INPUT][1];
cvTsPerspectiveTransform( test_array[INPUT][0], test_array[REF_OUTPUT][0], transmat );
}
CxCore_PerspectiveTransformTest perspective_test;
///////////////// Mahalanobis /////////////////////
class CxCore_MahalanobisTest : public CxCore_MatrixTest
{
public:
CxCore_MahalanobisTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images );
int prepare_test_case( int test_case_idx );
void run_func();
void prepare_to_validation( int test_case_idx );
};
CxCore_MahalanobisTest::CxCore_MahalanobisTest() :
CxCore_MatrixTest( "matrix-mahalanobis", "cvMahalanobis", 3, 1, false, true, 1 )
{
test_case_count = 100;
test_array[TEMP].push(NULL);
test_array[TEMP].push(NULL);
test_array[TEMP].push(NULL);
}
void CxCore_MahalanobisTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
if( cvTsRandInt(rng) & 1 )
sizes[INPUT][0].width = sizes[INPUT][1].width = 1;
else
sizes[INPUT][0].height = sizes[INPUT][1].height = 1;
sizes[TEMP][0] = sizes[TEMP][1] = sizes[INPUT][0];
sizes[INPUT][2].width = sizes[INPUT][2].height = sizes[INPUT][0].width + sizes[INPUT][0].height - 1;
sizes[TEMP][2] = sizes[INPUT][2];
types[TEMP][0] = types[TEMP][1] = types[TEMP][2] = types[INPUT][0];
}
void CxCore_MahalanobisTest::get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types, CvSize** whole_sizes, bool* are_images )
{
CxCore_MatrixTest::get_timing_test_array_types_and_sizes( test_case_idx,
sizes, types, whole_sizes, are_images );
sizes[INPUT][0].height = sizes[INPUT][1].height = 1;
}
int CxCore_MahalanobisTest::prepare_test_case( int test_case_idx )
{
int code = CxCore_MatrixTest::prepare_test_case( test_case_idx );
if( code > 0 && ts->get_testing_mode() == CvTS::CORRECTNESS_CHECK_MODE )
{
// make sure that the inverted "covariation" matrix is symmetrix and positively defined.
cvTsGEMM( &test_mat[INPUT][2], &test_mat[INPUT][2], 1., 0, 0., &test_mat[TEMP][2], CV_GEMM_B_T );
cvTsCopy( &test_mat[TEMP][2], &test_mat[INPUT][2] );
}
return code;
}
void CxCore_MahalanobisTest::run_func()
{
*((CvScalar*)(test_mat[OUTPUT][0].data.db)) =
cvRealScalar(cvMahalanobis(test_array[INPUT][0], test_array[INPUT][1], test_array[INPUT][2]));
}
void CxCore_MahalanobisTest::prepare_to_validation( int )
{
cvTsAdd( &test_mat[INPUT][0], cvScalarAll(1.),
&test_mat[INPUT][1], cvScalarAll(-1.),
cvScalarAll(0.), &test_mat[TEMP][0], 0 );
if( test_mat[INPUT][0].rows == 1 )
cvTsGEMM( &test_mat[TEMP][0], &test_mat[INPUT][2], 1.,
0, 0., &test_mat[TEMP][1], 0 );
else
cvTsGEMM( &test_mat[INPUT][2], &test_mat[TEMP][0], 1.,
0, 0., &test_mat[TEMP][1], 0 );
*((CvScalar*)(test_mat[REF_OUTPUT][0].data.db)) =
cvRealScalar(sqrt(cvTsCrossCorr(&test_mat[TEMP][0], &test_mat[TEMP][1])));
}
CxCore_MahalanobisTest mahalanobis_test;
///////////////// covarmatrix /////////////////////
class CxCore_CovarMatrixTest : public CxCore_MatrixTest
{
public:
CxCore_CovarMatrixTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
int prepare_test_case( int test_case_idx );
void run_func();
void prepare_to_validation( int test_case_idx );
CvTestPtrVec temp_hdrs;
uchar* hdr_data;
int flags, t_flag, len, count;
bool are_images;
};
CxCore_CovarMatrixTest::CxCore_CovarMatrixTest() :
CxCore_MatrixTest( "matrix-covar", "cvCalcCovarMatrix", 1, 1, true, false, 1 ),
flags(0), t_flag(0), are_images(false)
{
test_case_count = 100;
test_array[INPUT_OUTPUT].push(NULL);
test_array[REF_INPUT_OUTPUT].push(NULL);
test_array[TEMP].push(NULL);
test_array[TEMP].push(NULL);
support_testing_modes = CvTS::CORRECTNESS_CHECK_MODE;
}
void CxCore_CovarMatrixTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
int bits = cvTsRandInt(rng);
int i, single_matrix;
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
flags = bits & (CV_COVAR_NORMAL | CV_COVAR_USE_AVG | CV_COVAR_SCALE | CV_COVAR_ROWS );
single_matrix = flags & CV_COVAR_ROWS;
t_flag = (bits & 256) != 0;
const int min_count = 2;
if( !t_flag )
{
len = sizes[INPUT][0].width;
count = sizes[INPUT][0].height;
count = MAX(count, min_count);
sizes[INPUT][0] = cvSize(len, count);
}
else
{
len = sizes[INPUT][0].height;
count = sizes[INPUT][0].width;
count = MAX(count, min_count);
sizes[INPUT][0] = cvSize(count, len);
}
if( single_matrix && t_flag )
flags = (flags & ~CV_COVAR_ROWS) | CV_COVAR_COLS;
if( CV_MAT_DEPTH(types[INPUT][0]) == CV_32S )
types[INPUT][0] = (types[INPUT][0] & ~CV_MAT_DEPTH_MASK) | CV_32F;
sizes[OUTPUT][0] = sizes[REF_OUTPUT][0] = flags & CV_COVAR_NORMAL ? cvSize(len,len) : cvSize(count,count);
sizes[INPUT_OUTPUT][0] = sizes[REF_INPUT_OUTPUT][0] = !t_flag ? cvSize(len,1) : cvSize(1,len);
sizes[TEMP][0] = sizes[INPUT][0];
types[INPUT_OUTPUT][0] = types[REF_INPUT_OUTPUT][0] =
types[OUTPUT][0] = types[REF_OUTPUT][0] = types[TEMP][0] =
CV_MAT_DEPTH(types[INPUT][0]) == CV_64F || (bits & 512) ? CV_64F : CV_32F;
are_images = (bits & 1024) != 0;
for( i = 0; i < (single_matrix ? 1 : count); i++ )
temp_hdrs.push(NULL);
}
int CxCore_CovarMatrixTest::prepare_test_case( int test_case_idx )
{
int code = CxCore_MatrixTest::prepare_test_case( test_case_idx );
if( code > 0 )
{
int i;
int single_matrix = flags & (CV_COVAR_ROWS|CV_COVAR_COLS);
int hdr_size = are_images ? sizeof(IplImage) : sizeof(CvMat);
hdr_data = (uchar*)cvAlloc( count*hdr_size );
if( single_matrix )
{
if( !are_images )
*((CvMat*)hdr_data) = test_mat[INPUT][0];
else
cvGetImage( &test_mat[INPUT][0], (IplImage*)hdr_data );
temp_hdrs[0] = hdr_data;
}
else
for( i = 0; i < count; i++ )
{
CvMat part;
void* ptr = hdr_data + i*hdr_size;
if( !t_flag )
cvGetRow( &test_mat[INPUT][0], &part, i );
else
cvGetCol( &test_mat[INPUT][0], &part, i );
if( !are_images )
*((CvMat*)ptr) = part;
else
cvGetImage( &part, (IplImage*)ptr );
temp_hdrs[i] = ptr;
}
}
return code;
}
void CxCore_CovarMatrixTest::run_func()
{
cvCalcCovarMatrix( (const void**)&temp_hdrs[0], count,
test_array[OUTPUT][0], test_array[INPUT_OUTPUT][0], flags );
}
void CxCore_CovarMatrixTest::prepare_to_validation( int )
{
CvMat* avg = &test_mat[REF_INPUT_OUTPUT][0];
double scale = 1.;
if( !(flags & CV_COVAR_USE_AVG) )
{
int i;
cvTsZero( avg );
for( i = 0; i < count; i++ )
{
CvMat stub, *vec = 0;
if( flags & CV_COVAR_ROWS )
vec = cvGetRow( temp_hdrs[0], &stub, i );
else if( flags & CV_COVAR_COLS )
vec = cvGetCol( temp_hdrs[0], &stub, i );
else
vec = cvGetMat( temp_hdrs[i], &stub );
cvTsAdd( avg, cvScalarAll(1.), vec,
cvScalarAll(1.), cvScalarAll(0.), avg, 0 );
}
cvTsAdd( avg, cvScalarAll(1./count), 0,
cvScalarAll(0.), cvScalarAll(0.), avg, 0 );
}
if( flags & CV_COVAR_SCALE )
{
scale = 1./count;
}
cvRepeat( avg, &test_mat[TEMP][0] );
cvTsAdd( &test_mat[INPUT][0], cvScalarAll(1.),
&test_mat[TEMP][0], cvScalarAll(-1.),
cvScalarAll(0.), &test_mat[TEMP][0], 0 );
cvTsGEMM( &test_mat[TEMP][0], &test_mat[TEMP][0],
scale, 0, 0., &test_mat[REF_OUTPUT][0],
t_flag ^ ((flags & CV_COVAR_NORMAL) != 0) ?
CV_GEMM_A_T : CV_GEMM_B_T );
cvFree( &hdr_data );
temp_hdrs.clear();
}
CxCore_CovarMatrixTest covarmatrix_test;
static void cvTsFloodWithZeros( CvMat* mat, CvRNG* rng )
{
int k, total = mat->rows*mat->cols;
int zero_total = cvTsRandInt(rng) % total;
assert( CV_MAT_TYPE(mat->type) == CV_32FC1 ||
CV_MAT_TYPE(mat->type) == CV_64FC1 );
for( k = 0; k < zero_total; k++ )
{
int i = cvTsRandInt(rng) % mat->rows;
int j = cvTsRandInt(rng) % mat->cols;
uchar* row = mat->data.ptr + mat->step*i;
if( CV_MAT_DEPTH(mat->type) == CV_32FC1 )
((float*)row)[j] = 0.f;
else
((double*)row)[j] = 0.;
}
}
///////////////// determinant /////////////////////
class CxCore_DetTest : public CxCore_MatrixTest
{
public:
CxCore_DetTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
double get_success_error_level( int test_case_idx, int i, int j );
void get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high );
int prepare_test_case( int test_case_idx );
void run_func();
void prepare_to_validation( int test_case_idx );
};
CxCore_DetTest::CxCore_DetTest() :
CxCore_MatrixTest( "matrix-det", "cvDet", 1, 1, false, true, 1 )
{
test_case_count = 100;
max_log_array_size = 7;
test_array[TEMP].push(NULL);
}
void CxCore_DetTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
sizes[INPUT][0].width = sizes[INPUT][0].height = sizes[INPUT][0].height;
sizes[TEMP][0] = sizes[INPUT][0];
types[TEMP][0] = CV_64FC1;
}
void CxCore_DetTest::get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high )
{
*low = cvScalarAll(-2.);
*high = cvScalarAll(2.);
}
double CxCore_DetTest::get_success_error_level( int /*test_case_idx*/, int /*i*/, int /*j*/ )
{
return CV_MAT_DEPTH(cvGetElemType(test_array[INPUT][0])) == CV_32F ? 1e-2 : 1e-5;
}
int CxCore_DetTest::prepare_test_case( int test_case_idx )
{
int code = CxCore_MatrixTest::prepare_test_case( test_case_idx );
if( code > 0 )
cvTsFloodWithZeros( &test_mat[INPUT][0], ts->get_rng() );
return code;
}
void CxCore_DetTest::run_func()
{
*((CvScalar*)(test_mat[OUTPUT][0].data.db)) = cvRealScalar(cvDet(test_array[INPUT][0]));
}
// LU method that chooses the optimal in a column pivot element
static double cvTsLU( CvMat* a, CvMat* b=NULL, CvMat* x=NULL, int* rank=0 )
{
int i, j, k, N = a->rows, N1 = a->cols, Nm = MIN(N, N1), step = a->step/sizeof(double);
int M = b ? b->cols : 0, b_step = b ? b->step/sizeof(double) : 0;
int x_step = x ? x->step/sizeof(double) : 0;
double *a0 = a->data.db, *b0 = b ? b->data.db : 0;
double *x0 = x ? x->data.db : 0;
double t, det = 1.;
assert( CV_MAT_TYPE(a->type) == CV_64FC1 &&
(!b || CV_ARE_TYPES_EQ(a,b)) && (!x || CV_ARE_TYPES_EQ(a,x)));
for( i = 0; i < Nm; i++ )
{
double max_val = fabs(a0[i*step + i]);
double *a1, *a2, *b1 = 0, *b2 = 0;
k = i;
for( j = i+1; j < N; j++ )
{
t = fabs(a0[j*step + i]);
if( max_val < t )
{
max_val = t;
k = j;
}
}
if( k != i )
{
for( j = i; j < N1; j++ )
CV_SWAP( a0[i*step + j], a0[k*step + j], t );
for( j = 0; j < M; j++ )
CV_SWAP( b0[i*b_step + j], b0[k*b_step + j], t );
det = -det;
}
if( max_val == 0 )
{
if( rank )
*rank = i;
return 0.;
}
a1 = a0 + i*step;
a2 = a1 + step;
b1 = b0 + i*b_step;
b2 = b1 + b_step;
for( j = i+1; j < N; j++, a2 += step, b2 += b_step )
{
t = a2[i]/a1[i];
for( k = i+1; k < N1; k++ )
a2[k] -= t*a1[k];
for( k = 0; k < M; k++ )
b2[k] -= t*b1[k];
}
det *= a1[i];
}
if( x )
{
assert( b );
for( i = N-1; i >= 0; i-- )
{
double* a1 = a0 + i*step;
double* b1 = b0 + i*b_step;
for( j = 0; j < M; j++ )
{
t = b1[j];
for( k = i+1; k < N1; k++ )
t -= a1[k]*x0[k*x_step + j];
x0[i*x_step + j] = t/a1[i];
}
}
}
if( rank )
*rank = i;
return det;
}
void CxCore_DetTest::prepare_to_validation( int )
{
if( !CV_ARE_TYPES_EQ( &test_mat[INPUT][0], &test_mat[TEMP][0] ))
cvTsConvert( &test_mat[INPUT][0], &test_mat[TEMP][0] );
else
cvTsCopy( &test_mat[INPUT][0], &test_mat[TEMP][0], 0 );
*((CvScalar*)(test_mat[REF_OUTPUT][0].data.db)) = cvRealScalar(cvTsLU(&test_mat[TEMP][0], 0, 0));
}
CxCore_DetTest det_test;
///////////////// invert /////////////////////
static const char* matrix_solve_invert_param_names[] = { "size", "method", "depth", 0 };
static const char* matrix_solve_invert_methods[] = { "LU", "SVD", 0 };
class CxCore_InvertTest : public CxCore_MatrixTest
{
public:
CxCore_InvertTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images );
int write_default_params( CvFileStorage* fs );
void print_timing_params( int test_case_idx, char* ptr, int params_left );
void get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high );
double get_success_error_level( int test_case_idx, int i, int j );
int prepare_test_case( int test_case_idx );
void run_func();
void prepare_to_validation( int test_case_idx );
int method, rank;
double result;
};
CxCore_InvertTest::CxCore_InvertTest() :
CxCore_MatrixTest( "matrix-invert", "cvInvert, cvSVD, cvSVBkSb", 1, 1, false, false, 1 ), method(0), rank(0), result(0.)
{
test_case_count = 100;
max_log_array_size = 7;
test_array[TEMP].push(NULL);
test_array[TEMP].push(NULL);
default_timing_param_names = matrix_solve_invert_param_names;
}
void CxCore_InvertTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
int bits = cvTsRandInt(rng);
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
int min_size = MIN( sizes[INPUT][0].width, sizes[INPUT][0].height );
if( (bits & 3) == 0 )
{
method = CV_SVD;
if( bits & 4 )
{
sizes[INPUT][0] = cvSize(min_size, min_size);
if( bits & 16 )
method = CV_CHOLESKY;
}
}
else
{
method = CV_LU;
sizes[INPUT][0] = cvSize(min_size, min_size);
}
sizes[TEMP][0].width = sizes[INPUT][0].height;
sizes[TEMP][0].height = sizes[INPUT][0].width;
sizes[TEMP][1] = sizes[INPUT][0];
types[TEMP][0] = types[INPUT][0];
types[TEMP][1] = CV_64FC1;
sizes[OUTPUT][0] = sizes[REF_OUTPUT][0] = cvSize(min_size, min_size);
}
void CxCore_InvertTest::get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images )
{
CxCore_MatrixTest::get_timing_test_array_types_and_sizes( test_case_idx,
sizes, types, whole_sizes, are_images );
const char* method_str = cvReadString( find_timing_param("method"), "LU" );
method = strcmp( method_str, "LU" ) == 0 ? CV_LU : CV_SVD;
}
int CxCore_InvertTest::write_default_params( CvFileStorage* fs )
{
int code = CxCore_MatrixTest::write_default_params(fs);
if( code < 0 || ts->get_testing_mode() != CvTS::TIMING_MODE )
return code;
write_string_list( fs, "method", matrix_solve_invert_methods );
return code;
}
void CxCore_InvertTest::print_timing_params( int test_case_idx, char* ptr, int params_left )
{
sprintf( ptr, "%s,", method == CV_LU ? "LU" : "SVD" );
ptr += strlen(ptr);
params_left--;
CxCore_MatrixTest::print_timing_params( test_case_idx, ptr, params_left );
}
double CxCore_InvertTest::get_success_error_level( int /*test_case_idx*/, int, int )
{
return CV_MAT_DEPTH(cvGetElemType(test_array[OUTPUT][0])) == CV_32F ? 1e-2 : 1e-6;
}
int CxCore_InvertTest::prepare_test_case( int test_case_idx )
{
int code = CxCore_MatrixTest::prepare_test_case( test_case_idx );
if( code > 0 )
{
cvTsFloodWithZeros( &test_mat[INPUT][0], ts->get_rng() );
if( method == CV_CHOLESKY )
{
cvTsGEMM( &test_mat[INPUT][0], &test_mat[INPUT][0], 1.,
0, 0., &test_mat[TEMP][0], CV_GEMM_B_T );
cvTsCopy( &test_mat[TEMP][0], &test_mat[INPUT][0] );
}
}
return code;
}
void CxCore_InvertTest::get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high )
{
*low = cvScalarAll(-1.);
*high = cvScalarAll(1.);
}
void CxCore_InvertTest::run_func()
{
result = cvInvert(test_array[INPUT][0], test_array[TEMP][0], method);
}
static double cvTsSVDet( CvMat* mat, double* ratio )
{
int type = CV_MAT_TYPE(mat->type);
int i, nm = MIN( mat->rows, mat->cols );
CvMat* w = cvCreateMat( nm, 1, type );
double det = 1.;
cvSVD( mat, w, 0, 0, 0 );
if( type == CV_32FC1 )
{
for( i = 0; i < nm; i++ )
det *= w->data.fl[i];
*ratio = w->data.fl[nm-1] < FLT_EPSILON ? FLT_MAX : w->data.fl[nm-1]/w->data.fl[0];
}
else
{
for( i = 0; i < nm; i++ )
det *= w->data.db[i];
*ratio = w->data.db[nm-1] < FLT_EPSILON ? DBL_MAX : w->data.db[nm-1]/w->data.db[0];
}
cvReleaseMat( &w );
return det;
}
void CxCore_InvertTest::prepare_to_validation( int )
{
CvMat* input = &test_mat[INPUT][0];
double ratio = 0, det = cvTsSVDet( input, &ratio );
double threshold = (CV_MAT_DEPTH(input->type) == CV_32F ? FLT_EPSILON : DBL_EPSILON)*1000;
if( CV_MAT_TYPE(input->type) == CV_32FC1 )
cvTsConvert( input, &test_mat[TEMP][1] );
else
cvTsCopy( input, &test_mat[TEMP][1], 0 );
if( det < threshold ||
((method == CV_LU || method == CV_CHOLESKY) && (result == 0 || ratio < threshold)) ||
((method == CV_SVD || method == CV_SVD_SYM) && result < threshold) )
{
cvTsZero( &test_mat[OUTPUT][0] );
cvTsZero( &test_mat[REF_OUTPUT][0] );
//cvTsAdd( 0, cvScalarAll(0.), 0, cvScalarAll(0.), cvScalarAll(fabs(det)>1e-3),
// &test_mat[REF_OUTPUT][0], 0 );
return;
}
if( input->rows >= input->cols )
cvTsGEMM( &test_mat[TEMP][0], input, 1., 0, 0., &test_mat[OUTPUT][0], 0 );
else
cvTsGEMM( input, &test_mat[TEMP][0], 1., 0, 0., &test_mat[OUTPUT][0], 0 );
cvTsSetIdentity( &test_mat[REF_OUTPUT][0], cvScalarAll(1.) );
}
CxCore_InvertTest invert_test;
///////////////// solve /////////////////////
class CxCore_SolveTest : public CxCore_MatrixTest
{
public:
CxCore_SolveTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images );
int write_default_params( CvFileStorage* fs );
void print_timing_params( int test_case_idx, char* ptr, int params_left );
void get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high );
double get_success_error_level( int test_case_idx, int i, int j );
int prepare_test_case( int test_case_idx );
void run_func();
void prepare_to_validation( int test_case_idx );
int method, rank;
double result;
};
CxCore_SolveTest::CxCore_SolveTest() :
CxCore_MatrixTest( "matrix-solve", "cvSolve, cvSVD, cvSVBkSb", 2, 1, false, false, 1 ), method(0), rank(0), result(0.)
{
test_case_count = 100;
max_log_array_size = 7;
test_array[TEMP].push(NULL);
test_array[TEMP].push(NULL);
default_timing_param_names = matrix_solve_invert_param_names;
}
void CxCore_SolveTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
int bits = cvTsRandInt(rng);
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
CvSize in_sz = sizes[INPUT][0];
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
sizes[INPUT][0] = in_sz;
int min_size = MIN( sizes[INPUT][0].width, sizes[INPUT][0].height );
if( (bits & 3) == 0 )
{
method = CV_SVD;
if( bits & 4 )
{
sizes[INPUT][0] = cvSize(min_size, min_size);
/*if( bits & 8 )
method = CV_SVD_SYM;*/
}
}
else
{
method = CV_LU;
sizes[INPUT][0] = cvSize(min_size, min_size);
}
sizes[INPUT][1].height = sizes[INPUT][0].height;
sizes[TEMP][0].width = sizes[INPUT][1].width;
sizes[TEMP][0].height = sizes[INPUT][0].width;
sizes[TEMP][1] = sizes[INPUT][0];
types[TEMP][0] = types[INPUT][0];
types[TEMP][1] = CV_64FC1;
sizes[OUTPUT][0] = sizes[REF_OUTPUT][0] = cvSize(sizes[INPUT][1].width, min_size);
}
void CxCore_SolveTest::get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images )
{
CxCore_MatrixTest::get_timing_test_array_types_and_sizes( test_case_idx,
sizes, types, whole_sizes, are_images );
const char* method_str = cvReadString( find_timing_param("method"), "LU" );
sizes[INPUT][1].width = sizes[TEMP][0].width = sizes[OUTPUT][0].width = sizes[REF_OUTPUT][0].width = 1;
method = strcmp( method_str, "LU" ) == 0 ? CV_LU : CV_SVD;
}
int CxCore_SolveTest::write_default_params( CvFileStorage* fs )
{
int code = CxCore_MatrixTest::write_default_params(fs);
if( code < 0 || ts->get_testing_mode() != CvTS::TIMING_MODE )
return code;
write_string_list( fs, "method", matrix_solve_invert_methods );
return code;
}
void CxCore_SolveTest::print_timing_params( int test_case_idx, char* ptr, int params_left )
{
sprintf( ptr, "%s,", method == CV_LU ? "LU" : "SVD" );
ptr += strlen(ptr);
params_left--;
CxCore_MatrixTest::print_timing_params( test_case_idx, ptr, params_left );
}
int CxCore_SolveTest::prepare_test_case( int test_case_idx )
{
int code = CxCore_MatrixTest::prepare_test_case( test_case_idx );
/*if( method == CV_SVD_SYM )
{
cvTsGEMM( test_array[INPUT][0], test_array[INPUT][0], 1.,
0, 0., test_array[TEMP][0], CV_GEMM_B_T );
cvTsCopy( test_array[TEMP][0], test_array[INPUT][0] );
}*/
return code;
}
void CxCore_SolveTest::get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high )
{
*low = cvScalarAll(-1.);
*high = cvScalarAll(1.);
}
double CxCore_SolveTest::get_success_error_level( int /*test_case_idx*/, int, int )
{
return CV_MAT_DEPTH(cvGetElemType(test_array[OUTPUT][0])) == CV_32F ? 5e-2 : 1e-8;
}
void CxCore_SolveTest::run_func()
{
result = cvSolve(test_array[INPUT][0], test_array[INPUT][1], test_array[TEMP][0], method);
}
void CxCore_SolveTest::prepare_to_validation( int )
{
//int rank = test_mat[REF_OUTPUT][0].rows;
CvMat* dst;
CvMat* input = &test_mat[INPUT][0];
if( method == CV_LU )
{
if( result == 0 )
{
if( CV_MAT_TYPE(input->type) == CV_32FC1 )
cvTsConvert( input, &test_mat[TEMP][1] );
else
cvTsCopy( input, &test_mat[TEMP][1], 0 );
cvTsZero( &test_mat[OUTPUT][0] );
double det = cvTsLU( &test_mat[TEMP][1], 0, 0 );
cvTsAdd( 0, cvScalarAll(0.), 0, cvScalarAll(0.), cvScalarAll(det != 0),
&test_mat[REF_OUTPUT][0], 0 );
return;
}
double threshold = (CV_MAT_DEPTH(input->type) == CV_32F ? FLT_EPSILON : DBL_EPSILON)*1000;
double ratio = 0, det = cvTsSVDet( input, &ratio );
if( det < threshold || ratio < threshold )
{
cvTsZero( &test_mat[OUTPUT][0] );
cvTsZero( &test_mat[REF_OUTPUT][0] );
return;
}
}
dst = input->rows <= input->cols ? &test_mat[OUTPUT][0] : &test_mat[INPUT][1];
cvTsGEMM( input, &test_mat[TEMP][0], 1., &test_mat[INPUT][1], -1., dst, 0 );
if( dst != &test_mat[OUTPUT][0] )
cvTsGEMM( input, dst, 1., 0, 0., &test_mat[OUTPUT][0], CV_GEMM_A_T );
cvTsZero( &test_mat[REF_OUTPUT][0] );
}
CxCore_SolveTest solve_test;
///////////////// SVD /////////////////////
static const char* matrix_svd_param_names[] = { "size", "output", "depth", 0 };
static const char* matrix_svd_output_modes[] = { "w", "all", 0 };
class CxCore_SVDTest : public CxCore_MatrixTest
{
public:
CxCore_SVDTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images );
double get_success_error_level( int test_case_idx, int i, int j );
int write_default_params( CvFileStorage* fs );
void print_timing_params( int test_case_idx, char* ptr, int params_left );
void get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high );
int prepare_test_case( int test_case_idx );
void run_func();
void prepare_to_validation( int test_case_idx );
int flags;
bool have_u, have_v, symmetric, compact, vector_w;
};
CxCore_SVDTest::CxCore_SVDTest() :
CxCore_MatrixTest( "matrix-svd", "cvSVD", 1, 4, false, false, 1 ),
flags(0), have_u(false), have_v(false), symmetric(false), compact(false), vector_w(false)
{
test_case_count = 100;
test_array[TEMP].push(NULL);
test_array[TEMP].push(NULL);
test_array[TEMP].push(NULL);
test_array[TEMP].push(NULL);
default_timing_param_names = matrix_svd_param_names;
}
void CxCore_SVDTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
int bits = cvTsRandInt(rng);
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
int min_size, i, m, n;
min_size = MIN( sizes[INPUT][0].width, sizes[INPUT][0].height );
flags = bits & (CV_SVD_MODIFY_A+CV_SVD_U_T+CV_SVD_V_T);
have_u = (bits & 8) != 0;
have_v = (bits & 16) != 0;
symmetric = (bits & 32) != 0;
compact = (bits & 64) != 0;
vector_w = (bits & 128) != 0;
if( symmetric )
sizes[INPUT][0] = cvSize(min_size, min_size);
m = sizes[INPUT][0].height;
n = sizes[INPUT][0].width;
if( compact )
sizes[TEMP][0] = cvSize(min_size, min_size);
else
sizes[TEMP][0] = sizes[INPUT][0];
sizes[TEMP][3] = cvSize(0,0);
if( vector_w )
{
sizes[TEMP][3] = sizes[TEMP][0];
if( bits & 256 )
sizes[TEMP][0] = cvSize(1, min_size);
else
sizes[TEMP][0] = cvSize(min_size, 1);
}
if( have_u )
{
sizes[TEMP][1] = compact ? cvSize(min_size, m) : cvSize(m, m);
if( flags & CV_SVD_U_T )
CV_SWAP( sizes[TEMP][1].width, sizes[TEMP][1].height, i );
}
else
sizes[TEMP][1] = cvSize(0,0);
if( have_v )
{
sizes[TEMP][2] = compact ? cvSize(n, min_size) : cvSize(n, n);
if( !(flags & CV_SVD_V_T) )
CV_SWAP( sizes[TEMP][2].width, sizes[TEMP][2].height, i );
}
else
sizes[TEMP][2] = cvSize(0,0);
types[TEMP][0] = types[TEMP][1] = types[TEMP][2] = types[TEMP][3] = types[INPUT][0];
types[OUTPUT][0] = types[OUTPUT][1] = types[OUTPUT][2] = types[INPUT][0];
types[OUTPUT][3] = CV_8UC1;
sizes[OUTPUT][0] = !have_u || !have_v ? cvSize(0,0) : sizes[INPUT][0];
sizes[OUTPUT][1] = !have_u ? cvSize(0,0) : compact ? cvSize(min_size,min_size) : cvSize(m,m);
sizes[OUTPUT][2] = !have_v ? cvSize(0,0) : compact ? cvSize(min_size,min_size) : cvSize(n,n);
sizes[OUTPUT][3] = cvSize(min_size,1);
for( i = 0; i < 4; i++ )
{
sizes[REF_OUTPUT][i] = sizes[OUTPUT][i];
types[REF_OUTPUT][i] = types[OUTPUT][i];
}
}
void CxCore_SVDTest::get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images )
{
CxCore_MatrixTest::get_timing_test_array_types_and_sizes( test_case_idx,
sizes, types, whole_sizes, are_images );
const char* output_str = cvReadString( find_timing_param("output"), "all" );
bool need_all = strcmp( output_str, "all" ) == 0;
int i, count = test_array[OUTPUT].size();
vector_w = true;
symmetric = false;
compact = true;
sizes[TEMP][0] = cvSize(1,sizes[INPUT][0].height);
if( need_all )
{
have_u = have_v = true;
}
else
{
have_u = have_v = false;
sizes[TEMP][1] = sizes[TEMP][2] = cvSize(0,0);
}
flags = CV_SVD_U_T + CV_SVD_V_T;
for( i = 0; i < count; i++ )
sizes[OUTPUT][i] = sizes[REF_OUTPUT][i] = cvSize(0,0);
sizes[OUTPUT][0] = cvSize(1,1);
}
int CxCore_SVDTest::write_default_params( CvFileStorage* fs )
{
int code = CxCore_MatrixTest::write_default_params(fs);
if( code < 0 || ts->get_testing_mode() != CvTS::TIMING_MODE )
return code;
write_string_list( fs, "output", matrix_svd_output_modes );
return code;
}
void CxCore_SVDTest::print_timing_params( int test_case_idx, char* ptr, int params_left )
{
sprintf( ptr, "%s,", have_u ? "all" : "w" );
ptr += strlen(ptr);
params_left--;
CxCore_MatrixTest::print_timing_params( test_case_idx, ptr, params_left );
}
int CxCore_SVDTest::prepare_test_case( int test_case_idx )
{
int code = CxCore_MatrixTest::prepare_test_case( test_case_idx );
if( code > 0 )
{
CvMat* input = &test_mat[INPUT][0];
cvTsFloodWithZeros( input, ts->get_rng() );
if( symmetric && (have_u || have_v) )
{
CvMat* temp = &test_mat[TEMP][have_u ? 1 : 2];
cvTsGEMM( input, input, 1.,
0, 0., temp, CV_GEMM_B_T );
cvTsCopy( temp, input );
}
if( (flags & CV_SVD_MODIFY_A) && test_array[OUTPUT][0] )
cvTsCopy( input, &test_mat[OUTPUT][0] );
}
return code;
}
void CxCore_SVDTest::get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high )
{
*low = cvScalarAll(-2.);
*high = cvScalarAll(2.);
}
double CxCore_SVDTest::get_success_error_level( int test_case_idx, int i, int j )
{
int input_depth = CV_MAT_DEPTH(cvGetElemType( test_array[INPUT][0] ));
double input_precision = input_depth < CV_32F ? 0 : input_depth == CV_32F ?
5e-5 : 5e-11;
double output_precision = CvArrTest::get_success_error_level( test_case_idx, i, j );
return MAX(input_precision, output_precision);
}
void CxCore_SVDTest::run_func()
{
CvArr* src = test_array[!(flags & CV_SVD_MODIFY_A) ? INPUT : OUTPUT][0];
if( !src )
src = test_array[INPUT][0];
cvSVD( src, test_array[TEMP][0], test_array[TEMP][1], test_array[TEMP][2], flags );
}
void CxCore_SVDTest::prepare_to_validation( int )
{
CvMat* input = &test_mat[INPUT][0];
int m = input->rows, n = input->cols, min_size = MIN(m, n);
CvMat *src, *dst, *w;
double prev = 0, threshold = CV_MAT_TYPE(input->type) == CV_32FC1 ? FLT_EPSILON : DBL_EPSILON;
int i, j = 0, step;
if( have_u )
{
src = &test_mat[TEMP][1];
dst = &test_mat[OUTPUT][1];
cvTsGEMM( src, src, 1., 0, 0., dst, src->rows == dst->rows ? CV_GEMM_B_T : CV_GEMM_A_T );
cvTsSetIdentity( &test_mat[REF_OUTPUT][1], cvScalarAll(1.) );
}
if( have_v )
{
src = &test_mat[TEMP][2];
dst = &test_mat[OUTPUT][2];
cvTsGEMM( src, src, 1., 0, 0., dst, src->rows == dst->rows ? CV_GEMM_B_T : CV_GEMM_A_T );
cvTsSetIdentity( &test_mat[REF_OUTPUT][2], cvScalarAll(1.) );
}
w = &test_mat[TEMP][0];
step = w->rows == 1 ? 1 : w->step/CV_ELEM_SIZE(w->type);
for( i = 0; i < min_size; i++ )
{
double norm = 0, aii;
uchar* row_ptr;
if( w->rows > 1 && w->cols > 1 )
{
CvMat row;
cvGetRow( w, &row, i );
norm = cvNorm( &row, 0, CV_L1 );
j = i;
row_ptr = row.data.ptr;
}
else
{
row_ptr = w->data.ptr;
j = i*step;
}
aii = CV_MAT_TYPE(w->type) == CV_32FC1 ?
(double)((float*)row_ptr)[j] : ((double*)row_ptr)[j];
if( w->rows == 1 || w->cols == 1 )
norm = aii;
norm = fabs(norm - aii);
test_mat[OUTPUT][3].data.ptr[i] = aii >= 0 && norm < threshold && (i == 0 || aii <= prev);
prev = aii;
}
cvTsAdd( 0, cvScalarAll(0.), 0, cvScalarAll(0.),
cvScalarAll(1.), &test_mat[REF_OUTPUT][3], 0 );
if( have_u && have_v )
{
if( vector_w )
{
cvTsZero( &test_mat[TEMP][3] );
for( i = 0; i < min_size; i++ )
{
double val = cvGetReal1D( w, i );
cvSetReal2D( &test_mat[TEMP][3], i, i, val );
}
w = &test_mat[TEMP][3];
}
if( m >= n )
{
cvTsGEMM( &test_mat[TEMP][1], w, 1., 0, 0., &test_mat[REF_OUTPUT][0],
flags & CV_SVD_U_T ? CV_GEMM_A_T : 0 );
cvTsGEMM( &test_mat[REF_OUTPUT][0], &test_mat[TEMP][2], 1., 0, 0.,
&test_mat[OUTPUT][0], flags & CV_SVD_V_T ? 0 : CV_GEMM_B_T );
}
else
{
cvTsGEMM( w, &test_mat[TEMP][2], 1., 0, 0., &test_mat[REF_OUTPUT][0],
flags & CV_SVD_V_T ? 0 : CV_GEMM_B_T );
cvTsGEMM( &test_mat[TEMP][1], &test_mat[REF_OUTPUT][0], 1., 0, 0.,
&test_mat[OUTPUT][0], flags & CV_SVD_U_T ? CV_GEMM_A_T : 0 );
}
cvTsCopy( &test_mat[INPUT][0], &test_mat[REF_OUTPUT][0], 0 );
}
}
CxCore_SVDTest svd_test;
///////////////// SVBkSb /////////////////////
class CxCore_SVBkSbTest : public CxCore_MatrixTest
{
public:
CxCore_SVBkSbTest();
protected:
void get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types );
void get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images );
double get_success_error_level( int test_case_idx, int i, int j );
void get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high );
int prepare_test_case( int test_case_idx );
void run_func();
void prepare_to_validation( int test_case_idx );
int flags;
bool have_b, symmetric, compact, vector_w;
};
CxCore_SVBkSbTest::CxCore_SVBkSbTest() :
CxCore_MatrixTest( "matrix-svbksb", "cvSVBkSb", 2, 1, false, false, 1 ),
flags(0), have_b(false), symmetric(false), compact(false), vector_w(false)
{
test_case_count = 100;
test_array[TEMP].push(NULL);
test_array[TEMP].push(NULL);
test_array[TEMP].push(NULL);
}
void CxCore_SVBkSbTest::get_test_array_types_and_sizes( int test_case_idx, CvSize** sizes, int** types )
{
CvRNG* rng = ts->get_rng();
int bits = cvTsRandInt(rng);
CxCore_MatrixTest::get_test_array_types_and_sizes( test_case_idx, sizes, types );
int min_size, i, m, n;
CvSize b_size;
min_size = MIN( sizes[INPUT][0].width, sizes[INPUT][0].height );
flags = bits & (CV_SVD_MODIFY_A+CV_SVD_U_T+CV_SVD_V_T);
have_b = (bits & 16) != 0;
symmetric = (bits & 32) != 0;
compact = (bits & 64) != 0;
vector_w = (bits & 128) != 0;
if( symmetric )
sizes[INPUT][0] = cvSize(min_size, min_size);
m = sizes[INPUT][0].height;
n = sizes[INPUT][0].width;
sizes[INPUT][1] = cvSize(0,0);
b_size = cvSize(m,m);
if( have_b )
{
sizes[INPUT][1].height = sizes[INPUT][0].height;
sizes[INPUT][1].width = cvTsRandInt(rng) % 100 + 1;
b_size = sizes[INPUT][1];
}
if( compact )
sizes[TEMP][0] = cvSize(min_size, min_size);
else
sizes[TEMP][0] = sizes[INPUT][0];
if( vector_w )
{
if( bits & 256 )
sizes[TEMP][0] = cvSize(1, min_size);
else
sizes[TEMP][0] = cvSize(min_size, 1);
}
sizes[TEMP][1] = compact ? cvSize(min_size, m) : cvSize(m, m);
if( flags & CV_SVD_U_T )
CV_SWAP( sizes[TEMP][1].width, sizes[TEMP][1].height, i );
sizes[TEMP][2] = compact ? cvSize(n, min_size) : cvSize(n, n);
if( !(flags & CV_SVD_V_T) )
CV_SWAP( sizes[TEMP][2].width, sizes[TEMP][2].height, i );
types[TEMP][0] = types[TEMP][1] = types[TEMP][2] = types[INPUT][0];
types[OUTPUT][0] = types[REF_OUTPUT][0] = types[INPUT][0];
sizes[OUTPUT][0] = sizes[REF_OUTPUT][0] = cvSize( b_size.width, n );
}
void CxCore_SVBkSbTest::get_timing_test_array_types_and_sizes( int test_case_idx,
CvSize** sizes, int** types,
CvSize** whole_sizes, bool* are_images )
{
CxCore_MatrixTest::get_timing_test_array_types_and_sizes( test_case_idx,
sizes, types, whole_sizes, are_images );
have_b = true;
vector_w = true;
compact = true;
sizes[TEMP][0] = cvSize(1,sizes[INPUT][0].height);
sizes[INPUT][1] = sizes[OUTPUT][0] = sizes[REF_OUTPUT][0] = cvSize(1,sizes[INPUT][0].height);
flags = CV_SVD_U_T + CV_SVD_V_T;
}
int CxCore_SVBkSbTest::prepare_test_case( int test_case_idx )
{
int code = CxCore_MatrixTest::prepare_test_case( test_case_idx );
if( code > 0 )
{
CvMat* input = &test_mat[INPUT][0];
cvTsFloodWithZeros( input, ts->get_rng() );
if( symmetric )
{
CvMat* temp = &test_mat[TEMP][1];
cvTsGEMM( input, input, 1., 0, 0., temp, CV_GEMM_B_T );
cvTsCopy( temp, input );
}
cvSVD( input, test_array[TEMP][0], test_array[TEMP][1], test_array[TEMP][2], flags );
}
return code;
}
void CxCore_SVBkSbTest::get_minmax_bounds( int /*i*/, int /*j*/, int /*type*/, CvScalar* low, CvScalar* high )
{
*low = cvScalarAll(-2.);
*high = cvScalarAll(2.);
}
double CxCore_SVBkSbTest::get_success_error_level( int /*test_case_idx*/, int /*i*/, int /*j*/ )
{
return CV_MAT_DEPTH(cvGetElemType(test_array[INPUT][0])) == CV_32F ? 1e-3 : 1e-7;
}
void CxCore_SVBkSbTest::run_func()
{
cvSVBkSb( test_array[TEMP][0], test_array[TEMP][1], test_array[TEMP][2],
test_array[INPUT][1], test_array[OUTPUT][0], flags );
}
void CxCore_SVBkSbTest::prepare_to_validation( int )
{
CvMat* input = &test_mat[INPUT][0];
int i, m = input->rows, n = input->cols, min_size = MIN(m, n), nb;
bool is_float = CV_MAT_DEPTH(input->type) == CV_32F;
CvSize w_size = compact ? cvSize(min_size,min_size) : cvSize(m,n);
CvMat* w = &test_mat[TEMP][0];
CvMat* wdb = cvCreateMat( w_size.height, w_size.width, CV_64FC1 );
// use exactly the same threshold as in icvSVD... ,
// so the changes in the library and here should be synchronized.
double threshold = cvSum( w ).val[0]*2*(is_float ? FLT_EPSILON : DBL_EPSILON);
CvMat *u, *v, *b, *t0, *t1;
cvTsZero(wdb);
for( i = 0; i < min_size; i++ )
{
double wii = vector_w ? cvGetReal1D(w,i) : cvGetReal2D(w,i,i);
cvSetReal2D( wdb, i, i, wii > threshold ? 1./wii : 0. );
}
u = &test_mat[TEMP][1];
v = &test_mat[TEMP][2];
b = 0;
nb = m;
if( test_array[INPUT][1] )
{
b = &test_mat[INPUT][1];
nb = b->cols;
}
if( is_float )
{
u = cvCreateMat( u->rows, u->cols, CV_64F );
cvTsConvert( &test_mat[TEMP][1], u );
if( b )
{
b = cvCreateMat( b->rows, b->cols, CV_64F );
cvTsConvert( &test_mat[INPUT][1], b );
}
}
t0 = cvCreateMat( wdb->cols, nb, CV_64F );
if( b )
cvTsGEMM( u, b, 1., 0, 0., t0, !(flags & CV_SVD_U_T) ? CV_GEMM_A_T : 0 );
else if( flags & CV_SVD_U_T )
cvTsCopy( u, t0 );
else
cvTsTranspose( u, t0 );
if( is_float )
{
cvReleaseMat( &b );
if( !symmetric )
{
cvReleaseMat( &u );
v = cvCreateMat( v->rows, v->cols, CV_64F );
}
else
{
v = u;
u = 0;
}
cvTsConvert( &test_mat[TEMP][2], v );
}
t1 = cvCreateMat( wdb->rows, nb, CV_64F );
cvTsGEMM( wdb, t0, 1, 0, 0, t1, 0 );
if( !is_float || !symmetric )
{
cvReleaseMat( &t0 );
t0 = !is_float ? &test_mat[REF_OUTPUT][0] : cvCreateMat( test_mat[REF_OUTPUT][0].rows, nb, CV_64F );
}
cvTsGEMM( v, t1, 1, 0, 0, t0, flags & CV_SVD_V_T ? CV_GEMM_A_T : 0 );
cvReleaseMat( &t1 );
if( t0 != &test_mat[REF_OUTPUT][0] )
{
cvTsConvert( t0, &test_mat[REF_OUTPUT][0] );
cvReleaseMat( &t0 );
}
if( v != &test_mat[TEMP][2] )
cvReleaseMat( &v );
cvReleaseMat( &wdb );
}
CxCore_SVBkSbTest svbksb_test;
// TODO: eigenvv, invsqrt, cbrt, fastarctan, (round, floor, ceil(?)),
/* End of file. */
Reference in New Issue View Git Blame Copy Permalink