Now that you have the ability to have the AI look at multiple frames and make corrections, would it be possible to make an AI model that repairs damaged frames? To be more specific, if I have a video where frame numbers 23-25 are damaged with corrupted digital pixelation, could we get a model that looks at the before and after frames and repairs those frames.
Making the model able to judge frame corruption is a difficult task. Your problem can be solved like this, manually modify the video to replace the damaged frame with the adjacent frame, then pass the result to Chronos and turn on “Replace Duplicate Frames”.
1 Like
Oh okay. I never thought of that. I will give that a try. Thank you!
