HyperAI超神经

Zero Shot Semantic Segmentation On Coco Stuff

评估指标

Inductive Setting hIoU
Transductive Setting hIoU

评测结果

各个模型在此基准测试上的表现结果

比较表格
模型名称Inductive Setting hIoUTransductive Setting hIoU
zegclip-towards-adapting-clip-for-zero-shot40.848.5
zero-shot-semantic-segmentation-with38.2-
context-aware-feature-generation-for-zero18.219.5
a-closer-look-at-self-training-for-zero-label-34.8
otseg-multi-prompt-sinkhorn-attention-for41.449.5
semantic-projection-network-for-zero-and-few14.030.3
denseclip-extract-free-dense-labels-from-clip-45.0
exploring-regional-clues-in-clip-for-zero41.249.7
mvp-seg-multi-view-prompt-learning-for-open-45.5
2112-1475736.341.5
decoupling-zero-shot-semantic-segmentation33.2-
otseg-multi-prompt-sinkhorn-attention-for41.549.8
freeseg-free-mask-from-interpretable-45.3
sign-spatial-information-incorporated20.9-
19060081715.016.2