Minor updates to training data

We updated the training data as we made minor changes to ensure consistency. These changes are as follows:

• We modified the script “split_video_in_frames.py” such that the frames generated from the videos are labeled “frame_xxxxxx.png” instead of “xxxxxx.png” as before.
• We modified the keypoint annotations in the files “Video_xx_Keypoints.json” in such a way that the frame numbers are filled with leading zeros to be consistent with the frame numbers of the generated individual images, i.e., instead of “frame_125”, for example, “frame_000125” is now specified.
• A header has been added to the files “Video_xx_Cuts.csv” which is consistent with the surgical phase specifications provided in “Video_xx_Phases.csv”.
• We added a small amount of information to the video metadata regarding copyright and other additional information.