This python scripts split gdc presentation format videos in a specified folder into seperate scenes.
python main.py [name of the folder which contains video files to split] {-o [name of the folder to store split scenes]}
- Scan video from the beginning to detect a presentation region.
- If a presentation region is detected, clipping begins.
- The script tries to find the presentation region for every frame.
- If the region is not found, the script considers it as full screen scene and captures following sequence as an mp4 file.
- If the region is found, the script calculates the difference between frames.
- If the difference is small, the script considers it as a static image. (clipped into a jpg file)
- If the difference is large enough, the script consider it as transitioning or animating scene. (baked into a gif file)
argparse https://docs.python.org/3/library/argparse.html
imageio https://imageio.github.io/
numpy https://www.numpy.org/
opencv2 https://opencv.org/