ZoomNAS: Searching for Whole-body Human Pose Estimation in the Wild
This work addresses the problem of accurate and efficient whole-body pose estimation for computer vision applications, representing an incremental advance by combining NAS with a new dataset.
The paper tackles 2D whole-body human pose estimation by proposing ZoomNAS, a neural architecture search framework that jointly searches model architecture and connections, and introduces COCO-WholeBody V1.0, a large-scale dataset with 133 keypoints. Experiments show ZoomNAS achieves state-of-the-art results, with improvements such as 1.2% AP gain over previous methods on the new dataset.
This paper investigates the task of 2D whole-body human pose estimation, which aims to localize dense landmarks on the entire human body including body, feet, face, and hands. We propose a single-network approach, termed ZoomNet, to take into account the hierarchical structure of the full human body and solve the scale variation of different body parts. We further propose a neural architecture search framework, termed ZoomNAS, to promote both the accuracy and efficiency of whole-body pose estimation. ZoomNAS jointly searches the model architecture and the connections between different sub-modules, and automatically allocates computational complexity for searched sub-modules. To train and evaluate ZoomNAS, we introduce the first large-scale 2D human whole-body dataset, namely COCO-WholeBody V1.0, which annotates 133 keypoints for in-the-wild images. Extensive experiments demonstrate the effectiveness of ZoomNAS and the significance of COCO-WholeBody V1.0.