Unsupervised Person Re Identification On 12

Métriques

Rank-1

Rank-10

Rank-5

mAP

Résultats

Résultats de performance de divers modèles sur ce benchmark

					Paper Title
PCL-CLIP (O2CAP)	84.9	94.0	92.0	65.5	Prototypical Contrastive Learning-based CLIP Fine-tuning for Object Re-identification
TMGF	83.3	92.1	90.2	58.2	Transformer Based Multi-Grained Features for Unsupervised Person Re-Identification
PCL-CLIP (CAP)	79.0	91.1	88.4	53.6	Prototypical Contrastive Learning-based CLIP Fine-tuning for Object Re-identification
PCL-CLIP (CC)	77.9	87.2	85.2	56.4	Prototypical Contrastive Learning-based CLIP Fine-tuning for Object Re-identification
CA-Jaccard	75.1	87.3	84.3	44.3	CA-Jaccard: Camera-aware Jaccard Distance for Person Re-identification
TransReID-SSL (ViTi-S)	75	-	-	50.6	Self-Supervised Pre-Training for Transformer-Based Person Re-Identification
PPLR	73.3	86.5	83.5	42.2	Part-based Pseudo Label Refinement for Unsupervised Person Re-identification
VAL-PAT	67.5	-	-	38.9	Learning Transferable Pedestrian Representation from Multimodal Information Supervision
TransReID-SSL (ViT-S)	66.4	-	-	40.9	Self-Supervised Pre-Training for Transformer-Based Person Re-Identification
Cluster Contrast	62	76.7	71.8	33	Cluster Contrast for Unsupervised Person Re-Identification
Group Sampling	56.2	71.5	67.3	24.6	Rethinking Sampling Strategies for Unsupervised Person Re-identification
CACL	48.9	66.4	61.2	23	Cluster-guided Asymmetric Contrastive Learning for Unsupervised Person Re-Identification

0 of 12 row(s) selected.

Unsupervised Person Re Identification On 12 | SOTA | HyperAI