HyperAI
AccueilActualitésArticles de recherche récentsTutorielsEnsembles de donnéesWikiSOTAModèles LLMClassement GPUÉvénements
Recherche
À propos
HyperAI

Unsupervised Semantic Segmentation With 8

Métriques

mIoU

Résultats

Résultats de performance de divers modèles sur ce benchmark

Nom du modèle
mIoU
Paper TitleRepository
MaskCLIP26.4Extract Free Dense Labels from CLIP
TTD (MaskCLIP)31.0TTD: Text-Tag Self-Distillation Enhancing Image-Text Alignment in CLIP to Alleviate Single Tag Bias
TagAlign37.6TagAlign: Improving Vision-Language Alignment with Multi-Tag Classification
TCL33.9Learning to Generate Text-grounded Mask for Open-world Semantic Segmentation from Only Image-Text Pairs
ReCo22.3ReCo: Retrieve and Co-segment for Zero-shot Transfer-
ProxyCLIP39.6ProxyCLIP: Proxy Attention Improves CLIP for Open-Vocabulary Segmentation
TTD (TCL)37.4TTD: Text-Tag Self-Distillation Enhancing Image-Text Alignment in CLIP to Alleviate Single Tag Bias
COSMOS ViT-B/1633.7COSMOS: Cross-Modality Self-Distillation for Vision Language Pre-training
GroupViT (RedCaps)23.4GroupViT: Semantic Segmentation Emerges from Text Supervision
Trident44.3Harnessing Vision Foundation Models for High-Performance, Training-Free Open Vocabulary Segmentation
0 of 10 row(s) selected.