Skip to main content
. Author manuscript; available in PMC: 2022 Dec 20.
Published in final edited form as: Proc IEEE Comput Soc Conf Comput Vis Pattern Recognit. 2021 Nov 2;2021:2875–2884. doi: 10.1109/cvpr46437.2021.00290

Figure 1. Overview of our approach.

Figure 1.

Part 1: A typical behavior study starts with extraction of tracking data from videos. We show 7 keypoints for each mouse, and draw the trajectory of the nose keypoint. Part 2: Domain experts can either do data annotation (Classifier A) or task programming (Classifier B) to reduce classifier error. The middle panel shows annotated frames at 30Hz. Colors in the bottom plot represent interpolated performance based on classifier error at the circular markers (full results in Section 4.3). The size of the marker represents the error variance.