generic-library/opencv
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

correct the size of buf which holds integral sum/sumsq

Browse Source 
layer_dy is not initialized correctly, it results in a small height
of the buffer, and so memory write is out of range in cv::integral.
This commit is contained in:
Guo Yejun
2015-01-08 16:10:38 +08:00
parent df57d038b8
commit 595ec6b852
1 changed files with 6 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
modules/objdetect/src/cascadedetect.cpp
View File

@@ -449,6 +449,12 @@ bool FeatureEvaluator::updateScaleData( Size imgsz, const std::vector<float>& _s
s.ystep = sc >= 2 ? 1 : 2; s.ystep = sc >= 2 ? 1 : 2;
s.scale = sc; s.scale = sc;
s.szi = Size(sz.width+1, sz.height+1); s.szi = Size(sz.width+1, sz.height+1);
if( i == 0 )
{
layer_dy = s.szi.height;
}
if( layer_ofs.x + s.szi.width > sbufSize.width ) if( layer_ofs.x + s.szi.width > sbufSize.width )
{ {
layer_ofs = Point(0, layer_ofs.y + layer_dy); layer_ofs = Point(0, layer_ofs.y + layer_dy);