全局学习扩展的可见光-红外行人重识别

郭子强,肖璇,陶浩然,王少荣

Global learning-expanded visible-infrared person re-identification

Ziqiang GUO,Xuan XIAO,Haoran TAO,Shaorong WANG

表 3 LLCM数据集上GLE与先进方法的性能比较

Tab.3 Performance comparison of GLE and state-of-the-art methods on LLCM dataset

方法 可见光检索红外 红外检索可见光 R-1/% R-10/% R-20/% mAP/% R-1/% R-10/% R-20/% mAP/% DDAG[15] 48.0 79.2 86.1 52.3 40.3 71.4 79.6 48.4 AGW[7] 51.5 81.5 87.9 55.3 43.6 74.6 82.4 51.8 LbA[39] 50.8 84.3 91.1 55.6 43.8 78.2 86.6 53.1 CAJ[31] 56.5 85.3 90.9 59.8 48.8 79.5 85.3 56.6 DART[30] 60.4 87.1 91.9 63.2 52.2 80.7 87.0 59.8 MMN[33] 59.9 88.5 93.6 62.7 52.5 81.6 88.4 58.9 DEEN[16] 62.5 90.3 94.7 65.8 54.9 84.9 90.9 62.9 HOS-Net[35] 64.9 — — 67.9 56.4 — — 63.2 IDKL[18] 72.2 — — 66.4 70.7 — — 65.2 GLE 71.6 92.5 96.2 56.9 57.7 86.3 92.1 64.4