View-Invariant Probabilistic Embedding for Human Pose
This addresses the challenge of view-invariant pose analysis for vision algorithms in images and videos, representing an incremental improvement over existing methods.
The paper tackles the problem of recognizing similar human body poses across different camera views using only 2D joint keypoints, by proposing a probabilistic embedding approach that achieves higher accuracy in cross-view pose retrieval compared to 2D-to-3D pose lifting models.
Depictions of similar human body configurations can vary with changing viewpoints. Using only 2D information, we would like to enable vision algorithms to recognize similarity in human body poses across multiple views. This ability is useful for analyzing body movements and human behaviors in images and videos. In this paper, we propose an approach for learning a compact view-invariant embedding space from 2D joint keypoints alone, without explicitly predicting 3D poses. Since 2D poses are projected from 3D space, they have an inherent ambiguity, which is difficult to represent through a deterministic mapping. Hence, we use probabilistic embeddings to model this input uncertainty. Experimental results show that our embedding model achieves higher accuracy when retrieving similar poses across different camera views, in comparison with 2D-to-3D pose lifting models. We also demonstrate the effectiveness of applying our embeddings to view-invariant action recognition and video alignment. Our code is available at https://github.com/google-research/google-research/tree/master/poem.