Welcome!

As a start, here's the abstract from our proposal:
(might still change a bit)

Behavior Recognition is a challenge central to computer vision. With the increase of video data that has been seen over the past years, methods to automatically interpret this data and recognize behavior are of great importance. The Smart Vivarium provides an ideal environment for the development of automated behavior recognition. Although it doesn’t face some of the issues commonly seen in outside situations, it does provide a starting point from which to develop behavior recognition techniques. We propose an algorithm based on Dollar et al.'s behavior recognition. We modify this algorithm to incorporate relative positioning for each cuboid in a manner similar to that presented by Aagarwal et al. We also propose including displacement as an indication of behavior. Finally, we will use a multi-class classifier to label each instance of behavior.