The video for computing homographies.txt should be the orignal one or the resized one?

[sorata118]

According to the 2th step of processing our own video, we should resize the video to 256x448 and save the frames in my_video/rgb. But at the 6th step, register the video and save the computed homographies in my_video/homographies.txt. The first line in omnimatte/datasets/tennis/homographies.txt is size: 854 480. So I'm little bit confused which video should we use for computing the homograpyies? The one before resized or after resized?

[erikalu]

The method we used to compute homographies was more accurate for higher resolution images, which is why we added the option to use homographies computed on a different resolution. You can use any size so long as it is included in the first line of homographies.txt