Abstract
The study of animal behavior increasingly relies on (semi-) automatic methodsfor the extraction of relevant behavioral features from video or picture data.To date, several specialized software products exist to detect and trackanimals' positions in simple (laboratory) environments. Tracking animals intheir natural environments, however, often requires substantial customizationof the image processing algorithms to the problem-specific imagecharacteristics. Here we introduce BioTracker, an open-source computer visionframework, that provides programmers with core functionalities that areessential parts of a tracking software, such as video I/O, graphics overlaysand mouse and keyboard interfaces. BioTracker additionally provides a number ofdifferent tracking algorithms suitable for a variety of image recordingconditions. The main feature of BioTracker is however the straightforwardimplementation of new problem-specific tracking modules and vision algorithmsthat can build upon BioTracker's core functionalities. With this open-sourceframework the scientific community can accelerate their research and focus onthe development of new vision algorithms.