SmooSeg (DINO, ViT-S/8)

Reported on 6 benchmarks across 3 tasks

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Medical2 results

Semantic SegmentationonCOCO-Stuff-27
Clustering [Accuracy]
63.2
best: 81.1 (DynaSeg - FSF (ResNet-18 FPN))
Semantic SegmentationonCOCO-Stuff-27
Clustering [mIoU]
26.7
best: 54.1 (DynaSeg - FSF (ResNet-18 FPN))

Computer Vision2 results

Unsupervised Semantic SegmentationonCOCO-Stuff-27
Clustering [Accuracy]
63.2
best: 81.1 (DynaSeg - FSF (ResNet-18 FPN))
Unsupervised Semantic SegmentationonCOCO-Stuff-27
Clustering [mIoU]
26.7
best: 54.1 (DynaSeg - FSF (ResNet-18 FPN))

Audio2 results

10-shot image generationonCOCO-Stuff-27
Clustering [Accuracy]
63.2
best: 81.1 (DynaSeg - FSF (ResNet-18 FPN))
10-shot image generationonCOCO-Stuff-27
Clustering [mIoU]
26.7
best: 54.1 (DynaSeg - FSF (ResNet-18 FPN))