Multiple Drosophila tracking with behavior classification

Behavior changes in model organisms provides information on how the organisms react to the environment or given stimuli. This knowledge helps scientist understand the process of life that can be used in term of medicine or disease threatening. In order to observe the behavior changes, computer vision is considered to be a superior tool for locomotion analysis as it can provide more precise and time efficient data in comparison with manual measurement. Moreover, a classification tool that can utilize the result from machine vision is also required. In this study, the usage of Drosophila tracking tool and behavior classifier is introduced. The fly tracking system can efficiently detect and track flies inside a closed arena. The result, which is in form of locomotion data such as position and velocity, is processed with a classifier tool to generate behavior label to each fly throughout the input video. As a result, the overall process not only provides insightful data, but also reduces time consumption to deal with the behavior analysis of Drosophila in comparison to the same process done manually.