Recode Khan Academy videos into a low-bandwidth vector format

by Azeem Ghumman for Foundation for Learning Equality

Khan Academy videos are essentially a series of pen-strokes on a simple background image. If we could accurately identify the operations being performed (like drawing, erasing) and the visual objects being drawn, we can convert the video into a vector-based representation which will contain almost the same amount of information but would require a very low storage space. We also need a system to create new videos using this vector-based approach and the playback capability.