Open Vocabulary Semantic Segmentation On

mIoU

평가 결과

이 벤치마크에서 각 모델의 성능 결과

		Paper Title
FC-CLIP	56.2	Convolutions Die Hard: Open-Vocabulary Segmentation with Single Frozen Convolutional CLIP
SimSeg	34.5	A Simple Baseline for Open-Vocabulary Semantic Segmentation with Pre-trained Vision-language Model
TTD (TCL)	32.0	TTD: Text-Tag Self-Distillation Enhancing Image-Text Alignment in CLIP to Alleviate Single Tag Bias
CLIP Surgery (CLIP without any fine-tuning)	31.4	A Closer Look at the Explainability of Contrastive Language-Image Pre-training
TTD (MaskCLIP)	27.0	TTD: Text-Tag Self-Distillation Enhancing Image-Text Alignment in CLIP to Alleviate Single Tag Bias

0 of 5 row(s) selected.