基于SMPL模态分解与嵌入融合的多模态步态识别

吴越,梁铮,高巍,杨茂达,赵培森,邓红霞,常媛媛

Multi-modal gait recognition based on SMPL model decomposition and embedding fusion

Yue WU,Zheng LIANG,Wei GAO,Maoda YANG,Peisen ZHAO,Hongxia DENG,Yuanyuan CHANG

表 3 不同方法在Gait3D数据集上识别性能的对比结果

Tab.3 Comparison results of different methods on Gait3D dataset

模态 方法 来源 Rank-1 Rank-5 mAP/% mINP/% 轮廓 GaitSet[5] AAAI2019 36.70 58.30 30.01 17.30 GaitPart[6] CVPR2020 28.20 47.60 21.58 12.36 GaitGL[8] ICCV2021 29.70 48.50 22.29 13.26 GaitGCI[9] CVPR2023 50.30 68.50 39.50 24.30 GaitBase[10] CVPR2023 64.20 79.50 54.51 36.36 DyGait[11] ICCV2023 66.30 80.80 56.40 37.30 骨骼 GaitGraph[13] ICIP2021 8.30 16.60 7.14 4.80 GPGait[14] ICCV2023 22.50 — — — 轮廓+ 骨骼/SMPL MSAFF[17] IJCB2023 48.10 66.60 38.45 23.49 GaitRef[18] IJCB2023 49.00 69.30 40.69 25.26 GaitSTR[19] T-BIOM2024 65.10 81.30 55.59 36.84 SMPLGait[20] CVPR2022 46.30 64.50 37.16 22.23 HybirdGait[21] AAAI2024 53.30 72.00 43.29 26.65 DFGait 本研究 70.40 85.00 61.04 41.27