Startseite Neuigkeiten Neueste Forschungsarbeiten Tutorials Datensätze Wiki SOTA LLM-Modelle GPU-Rangliste Veranstaltungen

Deutsch

Unsupervised Semantic Segmentation With 3

Metriken

mIoU

Ergebnisse

Leistungsergebnisse verschiedener Modelle zu diesem Benchmark

Modellname	mIoU	Paper Title	Repository
TCL	24.0	Learning to Generate Text-grounded Mask for Open-world Semantic Segmentation from Only Image-Text Pairs
TTD (MaskCLIP)	32.0	TTD: Text-Tag Self-Distillation Enhancing Image-Text Alignment in CLIP to Alleviate Single Tag Bias
ReCo+	24.2	ReCo: Retrieve and Co-segment for Zero-shot Transfer	-
CLIPpy ViT-B	18.1	Perceptual Grouping in Contrastive Vision-Language Models
MaskCLIP	10.0	Extract Free Dense Labels from CLIP
Trident	47.6	Harnessing Vision Foundation Models for High-Performance, Training-Free Open Vocabulary Segmentation
ReCo	19.3	ReCo: Retrieve and Co-segment for Zero-shot Transfer	-
TTD (TCL)	27.0	TTD: Text-Tag Self-Distillation Enhancing Image-Text Alignment in CLIP to Alleviate Single Tag Bias
COSMOS ViT-B/16	34.7	COSMOS: Cross-Modality Self-Distillation for Vision Language Pre-training
ProxyCLIP	42.0	ProxyCLIP: Proxy Attention Improves CLIP for Open-Vocabulary Segmentation
TagAlign	27.5	TagAlign: Improving Vision-Language Alignment with Multi-Tag Classification

0 of 11 row(s) selected.