Release time:2022-06-20  Hits:

• Indexed by:Essay collection
• First Author:Shiqi Li
• Correspondence Author:Xiang Xiang
• Journal:IAPR International Conference on Pattern Recognition (ICPR) 2022 Workshops
• Included Journals:EI
• Affiliation of Author(s):Springer Lecture Notes in Computer Science
• Place of Publication:Germany
• Discipline:Engineering
• First-Level Discipline:Computer Science and Technology
• Funded by:Project of Independent Innovation Research Fund of Huazhong University of Science and Technology (20
• Document Type:C
• Key Words:Pose estimation, 3D pose, human pose, lightweight, attention
• Date of Publication:2022-06-20
• Abstract:Recent research on human pose estimation exploits complex structures to improve performance on benchmark datasets, ignoring the resource overhead and inference speed when the model is actually deployed. In this paper, we lighten the computation cost and parameters of the deconvolution head network in SimpleBaseline and introduce an attention mechanism that utilizes original, inter-level, and intra-level information to intensify the accuracy. Additionally, we propose a novel loss function called heatmap weighting loss, which generates weights for each pixel on the heatmap that makes the model more focused on keypoints. Experiments demonstrate our method achieves a balance between performance, resource volume, and inference speed. Specifically, our method can achieve 65.3 AP score on COCO test-dev, while the inference speed is 55 FPS and 18 FPS on the mobile GPU and CPU, respectively.