In this study, we have established a novel behavioral paradigm in which flies are housed in a 7in x 7in x 1.5in open field with water and food provided. Within the glass chamber, we diffuse and change the light to simulate the day/night transition and control the temperature and air pressure to simulate different weather conditions. Flies are free to move anywhere they choose, including walking, flying, and interacting with other males and females.
There are three main challenges for our fly tracking problem: (1) the contrast between the flies and their surrounding background is low, making the automatic object detection hard; (2) the size of a fly (around 3 ~ 6 pixels) is small and the appearances of flies are very similar to each other, so we cannot extract rich feature descriptors on flies to build distinctive object models, making the appearance-based object tracking methods unsuitable here; (3) the flies can fly as fast as 1.7 meters/second, or 30 pixels/second in videos captured by a 120fps video camera with the resolution of 480 X 848 pixels. The motion blur caused by fast-motion makes the object detection hard. Furthermore, the displacement of a fast-moving fly between two consecutive frames is 5 or 10 times of its object size, challenging the data-association-based tracking methods that relies on good object detection performance and continuous motion.
A 30-minutes video (120fps, 480*848 pixels) is about 3G. We crop and
downsize two short clips from it below for the demonstration:
(right-click images for downloading videos. It may take a while for downloading if your internet is slow.)
Clip 1: 16 seconds (23M)
Clip 1: 25 seconds (34M)