Noisy-Correspondence Learning for Text-to-Image Person Re-identification

Benchmark Model Rank Results
text-based-person-retrieval-on-icfg-pedesRDE#3R@1: 67.68R@5: 82.47R@10: 87.36mAP: 40.06mINP: 7.87
text-based-person-retrieval-on-rstpreid-1RDE#5R@1: 65.35R@5: 83.95R@10: 89.90mAP: 50.88mINP: 28.08
text-based-person-retrieval-with-noisyRDE#1Rank-1: 74.46Rank 10: 93.63Rank-5: 89.42mAP: 66.13mINP: 49.66
text-based-person-retrieval-with-noisy-1RDE#1Rank 1: 66.54Rank-10: 86.70Rank-5: 81.70mAP: 39.08mINP: 7.55
text-based-person-retrieval-with-noisy-2RDE#1Rank 1: 64.45Rank 10: 90.00Rank 5: 83.50mAP: 49.78mINP: 27.43