激光与光电子学进展, 2020, 57 (2): 021006, 网络出版: 2020-01-03
基于视频的实时多人姿态估计方法 下载: 1445次
Real-Time Multi-Person Video-Based Pose Estimation
图像处理 多人姿态估计 空间变换网络 语义信息 姿态距离 image processing multi-person pose estimation spatial transformer network semantic information pose distance
摘要
针对图像和视频中多人姿态估计存在人体边界框定位不准确、困难关键点检测精度有待提高等问题,设计了一套基于自顶向下框架的实时多人姿态估计模型。首先将深度可分离卷积加入目标检测算法中,提高人体检测器运行速度;然后基于特征金字塔网络结合上下文语义信息,采用在线难例挖掘算法解决困难关键点检测精度低的问题;最后结合空间变换网络与姿态相似度计算,剔除冗余姿态,改善边界框定位准确性。本文提出模型在2017MS COCO Test-dev数据集上的平均检测精度比Mask R-CNN模型提升了14.84%,比RMPE模型提升了2.43%,帧频达到 22 frame/s。
Abstract
For multi-person pose estimation in images and videos, it is necessary to address the inaccurate positioning of the human-bounding box and improve the detection accuracy of hard keypoints. This paper designs a real-time multi-person pose-estimation model based on a top-down framework. First, depth-separable convolution is added to the target-detection algorithm to improve the running speed of the human detector; then, by combining the feature pyramid network with context-semantic information, the online hard-example mining algorithm is used to solve the problem of low detection accuracy at hard keypoints. Finally, combining the spatial-transformation network and pose-similarity calculation, the redundant pose is eliminated and the accuracy of the bounding-box positioning is improved. In this paper, the average detection precision of the proposed model on the 2017MS COCO Test-dev dataset is 14.84% higher than that of the Mask R-CNN model, and 2.43% higher than that of the RMPE model. The frame frequency is 22 frame·s -1.
闫芬婷, 王鹏, 吕志刚, 丁哲, 乔梦雨. 基于视频的实时多人姿态估计方法[J]. 激光与光电子学进展, 2020, 57(2): 021006. Yan Fenting, Wang Peng, Lü Zhigang, Ding Zhe, Qiao Mengyu. Real-Time Multi-Person Video-Based Pose Estimation[J]. Laser & Optoelectronics Progress, 2020, 57(2): 021006.