Video tracking is an established technique to follow a single animal in an experimental setup. With basic image processing it is possible to find the object in an image and determine its spatial position. Identification is not necessary because there is only one animal. Automatic tracking of multiple animals, however, requires a different approach. Tracking multiple individually identified animals can be achieved by marking the animals. One technique that is used for tracking two animals uses the difference in apparent size to identify the animals (available in EthoVision Pro). In practice, this technique can only be used reliably with two animals. In order to identify and track more than two animals, a technique known as color tracking can be used: animals are marked with different colors that enable individual recognition by the tracking system (available in EthoVision Color-Pro).

During the past year we have been working on improvement of the algorithms to achieve more accurate and robust tracking results. In particular, improvements have been made on the identification of the colors (a pattern classification problem) and extension of the user interface to give the user more control over the process of color assignment and tracking. Apart from enhanced color tracking software, this work has resulted in valuable knowledge on lighting techniques and other important variables in the color tracking process: what colors can be used, what are optimal initial settings on the frame grabber, etc.

The result of this work is a prototype of a new version of EthoVision Color-Pro, able to adequately identify and track at least eight animals. The new program will be demonstrated during the conference.