Janelia Automatic Animal Behavior Annotator (JAABA®)
Video of animal movement is a valuable measure of their behavior. Translating video into interpretable results if done manually would be extremely time-consuming. Furthermore, motion has several attributes including basic statistics, such as speed, to more intricate and complex movement patterns. Additionally, scientific experiments often produce unexpected behaviors that need careful examination toward developing new insights.
The Howard Hughes Medical Institute has developed a machine-vision based approach for quantifying animal behavior. The Janelia Automatic Animal Behavior Annotator (JAABA®) helps researchers quickly extract valuable information from videos. The researchers utilized a MatLab-based computer program and video capture to measure differences among 2,000 transgenic Drosophila lines from the Rubin GAL4 collection.
The application has two major components. First, a biologist describes the animal behavior by using an annotator tool. JAABA® has built a database of common behavior to help in this process. Labels are then manually added to tag specific behaviors such as walking, stopping, chasing, and jumping. This “training” input is then used by the software to automatically classify behaviors, thus saving researcher’s time. The JAABA® interface was designed to accommodate the wide variety of animal behavior. The software is user-friendly with no knowledge of machine learning needed.
- Intuitive interface that enables biologists to have access to the power of machine vision
- Allows researchers to create customized training data sets to precisely describe behavior
- New behavior categories can be easily created and applied to all transgenic lines
- The software can help more effectively understand animal behavior. The initial JAABA® software has been developed for the Drosophila line.
- HHMI Janelia has plans to develop machine-vision software also for larvae and mice.
Kabra M, Robie A, Rivera-Alba M, Hirokawa J, Branson S and Branson K (2012). Machine Vision Tools for Quantitatively Measuring Animal Behavior in Large Scale Experiments. Front. Behav. Neurosci. Conference Abstract: Tenth International Congress of Neuroethology. doi: 10.3389/conf.fnbeh.2012.27.00041
Free for Non-Profit Research and available for Commercial License
For inquiries, please reference: