PartFormer: Awakening Latent Diverse Representation from Vision Transformer for Object Re-Identification

Benchmark Model Rank Results
person-re-identification-on-occluded-dukemtmcPartFormer–Rank-1: 62.5mAP: 70.2