Unsupervised Semantic Segmentation on PascalVOC-20

Metric: mIoU (higher is better)

LeaderboardDataset

Loading chart...

Results

Submit a result

Sort:

#	Model↕	mIoU▼	Extra Data	Paper	Date↕	Code
1	CorrCLIP	91.8	No	CorrCLIP: Reconstructing Correlations in CLIP wi...	2024-11-15	Code
2	TextRegion	89.5	No	TextRegion: Text-Aligned Region Tokens from Froz...	2025-05-29	Code
3	Trident	88.7	No	Harnessing Vision Foundation Models for High-Per...	2024-11-14	Code
4	TagAlign	87.9	No	TagAlign: Improving Vision-Language Alignment wi...	2023-12-21	Code
5	ProxyCLIP	83.3	No	ProxyCLIP: Proxy Attention Improves CLIP for Ope...	2024-08-09	Code
6	TCL	83.2	No	Learning to Generate Text-grounded Mask for Open...	2022-12-01	Code
7	GroupViT (RedCaps)	79.7	No	GroupViT: Semantic Segmentation Emerges from Tex...	2022-02-22	Code
8	COSMOS ViT-B/16	77.7	No	COSMOS: Cross-Modality Self-Distillation for Vis...	2024-12-02	Code
9	MaskCLIP	74.9	No	Extract Free Dense Labels from CLIP	2021-12-02	Code
10	ReCo	57.7	No	ReCo: Retrieve and Co-segment for Zero-shot Tran...	2022-06-14	Code

#1CorrCLIPSOTA
91.8
mIoU· 2024-11-15
CorrCLIP: Reconstructing Correlations in CLIP with Off-the-Shelf Foundation Models for Open-Vocabulary Semantic Segmentation Code
#2TextRegion
89.5
mIoU· 2025-05-29
TextRegion: Text-Aligned Region Tokens from Frozen Image-Text Models Code
#3TridentSOTA
88.7
mIoU· 2024-11-14
Harnessing Vision Foundation Models for High-Performance, Training-Free Open Vocabulary Segmentation Code
#4TagAlignSOTA
87.9
mIoU· 2023-12-21
TagAlign: Improving Vision-Language Alignment with Multi-Tag Classification Code
#5ProxyCLIP
83.3
mIoU· 2024-08-09
ProxyCLIP: Proxy Attention Improves CLIP for Open-Vocabulary Segmentation Code
#6TCLSOTA
83.2
mIoU· 2022-12-01
Learning to Generate Text-grounded Mask for Open-world Semantic Segmentation from Only Image-Text Pairs Code
#7GroupViT (RedCaps)SOTA
79.7
mIoU· 2022-02-22
GroupViT: Semantic Segmentation Emerges from Text Supervision Code
#8COSMOS ViT-B/16
77.7
mIoU· 2024-12-02
COSMOS: Cross-Modality Self-Distillation for Vision Language Pre-training Code
#9MaskCLIPSOTA
74.9
mIoU· 2021-12-02
Extract Free Dense Labels from CLIP Code
#10ReCo
57.7
mIoU· 2022-06-14
ReCo: Retrieve and Co-segment for Zero-shot Transfer Code