Abstract: Highlights•The most comprehensive survey of text-based person re-identification to date.•A novel taxonomy for text-based person re-identification from different dimensions.•Challenges in text-based person re-ID are analyzed, with future directions given.•A open-world baseline for text-based pedestrian image generation guided Re-ID.
External IDs:doi:10.1016/j.inffus.2024.102879
Loading