Occluded person re-identification: a taxonomic survey and reproducible empirical benchmark

TL;DR

A unified empirical evaluation on the Market-1501, Occluded-DukeMTMC, and MSMT17 datasets is presented, revealing that latent feature-space refinement and semantic cross-modal alignment offer superior stability, scalability, and robustness compared to explicit pixel-level generation.