SOTAVerified

Unsupervised Semantic Segmentation with Language-image Pre-training

A segmentation task which does not utilise any human-level supervision for semantic segmentation except for a backbone which is initialised with features pre-trained with image-level labels.

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 1–14 of 14 papers

TitleStatusHype
Harnessing Vision Foundation Models for High-Performance, Training-Free Open Vocabulary SegmentationCode2
GroupViT: Semantic Segmentation Emerges from Text SupervisionCode2
ProxyCLIP: Proxy Attention Improves CLIP for Open-Vocabulary SegmentationCode2
TextRegion: Text-Aligned Region Tokens from Frozen Image-Text ModelsCode2
CorrCLIP: Reconstructing Correlations in CLIP with Off-the-Shelf Foundation Models for Open-Vocabulary Semantic SegmentationCode2
Perceptual Grouping in Contrastive Vision-Language ModelsCode1
ReCo: Retrieve and Co-segment for Zero-shot TransferCode1
TagAlign: Improving Vision-Language Alignment with Multi-Tag ClassificationCode1
TagCLIP: A Local-to-Global Framework to Enhance Open-Vocabulary Multi-Label Classification of CLIP Without TrainingCode1
TTD: Text-Tag Self-Distillation Enhancing Image-Text Alignment in CLIP to Alleviate Single Tag BiasCode1
COSMOS: Cross-Modality Self-Distillation for Vision Language Pre-trainingCode1
Extract Free Dense Labels from CLIPCode1
ITACLIP: Boosting Training-Free Semantic Segmentation with Image, Text, and Architectural EnhancementsCode1
Learning to Generate Text-grounded Mask for Open-world Semantic Segmentation from Only Image-Text PairsCode1
Show:102550

No leaderboard results yet.