|Goals||Technology||Community||Application Domains||User Studies|
VisionBlocks is a research project aimed at making computer vision accessible to masses. The primary goal of this project is to develop tools for delivering computer vision to masses. We present here an on-demand, in-browser and customizable computer vision application building system.
Can a grandma, with no prior programming experience, visit a website and build a computer vision application? Imagine a situation where you want to watch your baby when you are not in the room. Your friend wants to catch speeders in the neighborhood with his web-camera. Carol wants to track her favorite player in a football game. State-of-the-art computer vision algorithms can easily provide solutions in such situations. Yet, building such computer vision systems is beyond the reach of average consumer.
Advances in computer vision have traditionally been limited to a narrow circle of researchers. visionblocks is a web-based visual programming framework that empowers users with no programming experience to create robust and customizable vision systems. In visionblocks framework, users put objects together to build programs as if they were solving a puzzle. Our goal is to deliver computer vision to masses at right level of abstraction.