For metadata: it's given, you can use the "wild" method to do so. ROMP, VIBE and so on.
For masks: you use SAM, rmbeg or use the 3_ch_image function given in this repo.but for that you need to have your original image and a corresponding black image of the same size.
Extract frames: just use OPENCV to extract frames from videos.
For metadata: it's given, you can use the "wild" method to do so. ROMP, VIBE and so on.
For masks: you use SAM, rmbeg or use the 3_ch_image function given in this repo.but for that you need to have your original image and a corresponding black image of the same size.
Extract frames: just use OPENCV to extract frames from videos.