a nice-to-have would be the option (e.g. "--alignment-mask=maskfile.png") to supply a mask (image file) to be used in the alignment process, so the findTransformECC ignores parts of the images (inputMask parameter of cv::findTransformECC).
In my use case it would be due to images from a camera with extreme optical vignetting effect at the corners reducing quality of alignment.
Hi,
a nice-to-have would be the option (e.g. "--alignment-mask=maskfile.png") to supply a mask (image file) to be used in the alignment process, so the findTransformECC ignores parts of the images (inputMask parameter of cv::findTransformECC).
In my use case it would be due to images from a camera with extreme optical vignetting effect at the corners reducing quality of alignment.
Best regards David