Forums - What convolution is used in fcvConvValids16()?

This function does convolution of two images. Similar to 1-D convolution you flip one image and slide it through the other image to compute the output. Values are computed for the region where one image is completely within the other image. Following are details of parameters:

/// @param src1
///   First source image
///   \n\b WARNING: must be 128-bit aligned.
///
/// @param src1Width
///   Image width.
///
/// @param src1Height
///   Image height.
///
/// @param src1Stride
///   Stride of source image (i.e., how many bytes between column 0 of row 1 and
///   column 0 of row 2).
///   If set to 0, srcStride is srcWidth in bytes as default
///
/// @param src2
///   Second source image
///   \n\b WARNING: must be 128-bit aligned.
///
/// @param src2Width
///   Image width.
///   Must meet this condition: src2Width <= src1Width
///
/// @param src2Height
///   Image height.
///   Must meet this condition: src2Height <= src1Height
///
/// @param src2Stride
///   Stride of source images (i.e., how many bytes between column 0 of row 1 and
///   column 0 of row 2).
///   If set to 0, srcStride is src2Width in bytes as default
///
/// @param dst
///   Destination.
///   Size of destination is (src1Width-src2Width+1) x (src1Height-src2Height+1)
///   \n\b WARNING: must be 128-bit aligned.
///
/// @param dstStride
///   Stride of destination image (i.e., how many bytes between column 0 of row 1 and
///   column 0 of row 2).