diff --git a/modules/imgproc/include/opencv2/imgproc.hpp b/modules/imgproc/include/opencv2/imgproc.hpp
index a9fa7ae21..14f4067a9 100644
--- a/modules/imgproc/include/opencv2/imgproc.hpp
+++ b/modules/imgproc/include/opencv2/imgproc.hpp
@@ -326,7 +326,9 @@ enum AdaptiveThresholdTypes {
     window) of the \f$\texttt{blockSize} \times \texttt{blockSize}\f$ neighborhood of \f$(x, y)\f$
     minus C . The default sigma (standard deviation) is used for the specified blockSize . See
     cv::getGaussianKernel*/
-    ADAPTIVE_THRESH_GAUSSIAN_C = 1
+    ADAPTIVE_THRESH_GAUSSIAN_C = 1,
+     /** Like ADAPTIVE_THRESH_GAUSSIAN_C except that GaussianBlur use CV_32F for blurring to avoid rounding error*/
+   ADAPTIVE_THRESH_GAUSSIAN_C_FLOAT = 2
 };
 
 //! cv::undistort mode
diff --git a/modules/imgproc/src/thresh.cpp b/modules/imgproc/src/thresh.cpp
index 490bdff3e..0eae14546 100644
--- a/modules/imgproc/src/thresh.cpp
+++ b/modules/imgproc/src/thresh.cpp
@@ -1295,11 +1295,19 @@ void cv::adaptiveThreshold( InputArray _src, OutputArray _dst, double maxValue,
     if( src.data != dst.data )
         mean = dst;
 
-    if( method == ADAPTIVE_THRESH_MEAN_C )
+    if (method == ADAPTIVE_THRESH_MEAN_C)
         boxFilter( src, mean, src.type(), Size(blockSize, blockSize),
                    Point(-1,-1), true, BORDER_REPLICATE );
-    else if( method == ADAPTIVE_THRESH_GAUSSIAN_C )
-        GaussianBlur( src, mean, Size(blockSize, blockSize), 0, 0, BORDER_REPLICATE );
+    else if (method == ADAPTIVE_THRESH_GAUSSIAN_C)
+        GaussianBlur(src, mean, Size(blockSize, blockSize), 0, 0, BORDER_REPLICATE);
+    else if (method == ADAPTIVE_THRESH_GAUSSIAN_C_FLOAT)
+    {
+        Mat srcfloat,meanfloat;
+        src.convertTo(srcfloat,CV_32F);
+        meanfloat=srcfloat;
+        GaussianBlur(srcfloat, meanfloat, Size(blockSize, blockSize), 0, 0, BORDER_REPLICATE);
+        meanfloat.convertTo(dst, src.type());
+    }
     else
         CV_Error( CV_StsBadFlag, "Unknown/unsupported adaptive threshold method" );