MONET: Multiview Semi-supervised Keypoint Detection via Epipolar Divergence
This work addresses the problem of limited annotated data for keypoint detection in general subjects like animals, offering a semi-supervised solution that is incremental in improving geometry integration.
The paper tackles the challenge of semi-supervised keypoint detection for non-human species by introducing a new differentiable representation called epipolar divergence, which integrates multiview geometry into learning and achieves localization of keypoints across diverse species like humans, dogs, and monkeys.
This paper presents MONET -- an end-to-end semi-supervised learning framework for a keypoint detector using multiview image streams. In particular, we consider general subjects such as non-human species where attaining a large scale annotated dataset is challenging. While multiview geometry can be used to self-supervise the unlabeled data, integrating the geometry into learning a keypoint detector is challenging due to representation mismatch. We address this mismatch by formulating a new differentiable representation of the epipolar constraint called epipolar divergence---a generalized distance from the epipolar lines to the corresponding keypoint distribution. Epipolar divergence characterizes when two view keypoint distributions produce zero reprojection error. We design a twin network that minimizes the epipolar divergence through stereo rectification that can significantly alleviate computational complexity and sampling aliasing in training. We demonstrate that our framework can localize customized keypoints of diverse species, e.g., humans, dogs, and monkeys.