摘要
人体姿态估计是计算机视觉领域的热门研究课题。随着深度学习的发展,人体姿态估计模型已经能够精准预测人体关键点。针对关键点被遮挡、关键点重合以及复杂背景等问题,提出了一种结合注意力机制的级联金字塔模型,它将注意力机制加入特征提取网络中,使模型可以获得更丰富的特征信息,并且借助GlobalNet和RefineNet达到精准定位被遮挡关键点的目的。在公开数据集MPII、MS COCO2017和3DOH50K上的验证结果表明,相较于以往模型,该模型在标准情况和被遮挡情况下人体姿态估计的准确度有所提升,且具有鲁棒性。
Human pose estimation is a popular research topic in the field of computer vision.With the development of deep learning,human pose estimation models can accurately predict human key points.Aiming at the problems of occlusion of key points,overlapping of key points and complex background,this paper proposes a cascade pyramid model combined with attention mechanism.By integrating attention mechanism into the feature extraction network,this model can obtain richer feature information.With the help of GlobalNet and RefineNet,it can accurately locate the occluded key points.It is shown via the results on public dataset,MPII,MS COCO2017 and 3DOH50K,that this model can attain higher accuracy of human pose estimation in standard and occluded situations than previous models,and has better robustness.
作者
牛悦
王安南
吴胜昔
NIU Yue;WANG Annan;WU Shengxi(Key Laboratory of Smart Manufacturing in Energy Chemical Process,Ministry of Education,East China University of Science and Technology,Shanghai 200237,China)
出处
《华东理工大学学报(自然科学版)》
CAS
CSCD
北大核心
2023年第5期724-734,共11页
Journal of East China University of Science and Technology
基金
国家自然科学基金(61973120)
上海汽车工业科技发展基金(1837)。