Rethinking the Heatmap Regression for Bottom-up Human Pose Estimation | |
Luo, Zhengxiong1,2,3,4; Wang, Zhicheng1; Huang, Yan3,4; Wang, Liang3,4; Tan, Tieniu3; Zhou, Erjin1 | |
2021-06 | |
会议名称 | IEEE Conference of Computer Vision and Pattern Recognition |
会议日期 | 2021-6 |
会议地点 | 美国纳什维尔 |
摘要 | Heatmap regression has become the most prevalent choice for nowadays human pose estimation methods. The ground-truth heatmaps are usually constructed via cover- ing all skeletal keypoints by 2D gaussian kernels. The standard deviations of these kernels are fixed. However, for bottom-up methods, which need to handle a large variance of human scales and labeling ambiguities, the current practice seems unreasonable. To better cope with these problems, we propose the scale-adaptive heatmap regression (SAHR) method, which can adaptively adjust the standard deviation for each keypoint. In this way, SAHR is more tolerant of various human scales and labeling ambiguities. However, SAHR may aggravate the imbalance between fore-background samples, which potentially hurts the improvement of SAHR. Thus, we further introduce the weight-adaptive heatmap regression (WAHR) to help balance the fore-background samples. Extensive experiments show that SAHR together with WAHR largely improves the accuracy of bottom-up human pose estimation. As a result, we finally outperform the state-of-the-art model by +1.5AP and achieve 72.0AP on COCO test-dev2017, which is comparable with the performances of most top-down methods. Source codes are available at https://github.com/ greatlog/SWAHR-HumanPose. |
七大方向——子方向分类 | 目标检测、跟踪与识别 |
国重实验室规划方向分类 | 视觉信息处理 |
是否有论文关联数据集需要存交 | 否 |
文献类型 | 会议论文 |
条目标识符 | http://ir.ia.ac.cn/handle/173211/51942 |
专题 | 模式识别实验室 |
通讯作者 | Huang, Yan |
作者单位 | 1.Megvii Inc 2.University of Chinese Academy of Sciences (UCAS) 3.Center for Research on Intelligent Perception and Computing (CRIPAC), National Laboratory of Pattern Recognition (NLPR) 4.Institute of Automation, Chinese Academy of Sciences (CASIA) |
第一作者单位 | 模式识别国家重点实验室; 中国科学院自动化研究所 |
通讯作者单位 | 模式识别国家重点实验室; 中国科学院自动化研究所 |
推荐引用方式 GB/T 7714 | Luo, Zhengxiong,Wang, Zhicheng,Huang, Yan,et al. Rethinking the Heatmap Regression for Bottom-up Human Pose Estimation[C],2021. |
条目包含的文件 | ||||||
文件名称/大小 | 文献类型 | 版本类型 | 开放类型 | 使用许可 | ||
2021 - Rethinking th(2163KB) | 会议论文 | 开放获取 | CC BY-NC-SA | 浏览 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论