UnCageNet: Tracking and Pose Estimation of Caged Animal
This addresses a domain-specific problem for researchers in animal behavior and computer vision by providing an incremental solution to handle systematic occlusions in cage structures.
The paper tackled the problem of performance drops in animal tracking and pose estimation due to cage occlusions by developing a three-stage preprocessing pipeline, resulting in performance comparable to environments without occlusions and significant improvements in keypoint detection accuracy and trajectory consistency.
Animal tracking and pose estimation systems, such as STEP (Simultaneous Tracking and Pose Estimation) and ViTPose, experience substantial performance drops when processing images and videos with cage structures and systematic occlusions. We present a three-stage preprocessing pipeline that addresses this limitation through: (1) cage segmentation using a Gabor-enhanced ResNet-UNet architecture with tunable orientation filters, (2) cage inpainting using CRFill for content-aware reconstruction of occluded regions, and (3) evaluation of pose estimation and tracking on the uncaged frames. Our Gabor-enhanced segmentation model leverages orientation-aware features with 72 directional kernels to accurately identify and segment cage structures that severely impair the performance of existing methods. Experimental validation demonstrates that removing cage occlusions through our pipeline enables pose estimation and tracking performance comparable to that in environments without occlusions. We also observe significant improvements in keypoint detection accuracy and trajectory consistency.