Hello, your algorithm is very interesting, but I have a question, your output depth map is 8 bits png, not the 16-bit depth map, I noticed that the image pixels in the code are normalized to 0-255, so how to get the correct 16-bit depth map that has been completed
I created a visual studio project using the given header file and its .cpp file, and when I try to run the main file using images in the Example folder, Visual studio gave multiple run time errors saying:
"_Error: Assertion failed ((unsigned)(i1 * DataType<Tp>::channels) < (unsigned)(size.p[1] * channels())) in cv::Mat::at"
in the for loop of "preProcess"
and
" _Error: Assertion failed ((unsigned)(i1 * DataType<Tp>::channels) < (unsigned)(size.p[1] * channels())) in cv::Mat::at"
in the for loop of "identFillHoles".
Hi, how should I pass the segmented image to the algorithm? Is a binary mask enough if I have only 1 class on which I would like to do the depth completion?