Home

Overview

This project consists of 3 parts.

1. Vectorization - Vectorize currently existing blackboard style educational videos. This consists of identifying user actions, such as drawing and erasing, and extracting these actions into a time-stamped vector-based representation in a JSON format.
2. Rendering - Vector Video Player that takes a JSON file as input, and renders the vector video format within a browser using JavaScript and the HTML5 Canvas.
3. Authoring - Allow authoring, editing, and localizing vector videos within a browser.

Motivation

• Huge space/bandwidths savings, 250x or better.
• Support smaller screens using adaptive zoom.
• Allow authoring and editing of educational content without the need of screen capture and video encoding.
• Allow content to be created/edited/localized even on low-powered devices.

Status

1. Vectorization - Currently, we can process basic Khan Academy Videos. However, we are planning to handle more complicated features including panning, zooming, etc, and even videos that involve an instructor in front of a chalk/white board.

2. Rendering - A basic alpha version of the Vector Video Player has been developed and can be viewed in the repo.

3. Authoring - Coming soon.