WSI-level classification on Private-Breast
68Top-1 AccuracyPathDino-512
Evaluation Results
| Method | Links | |
|---|---|---|
| PathDino-512Pre-training domain=Histopathology, Architecture type=Transformer, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 68 | |
| iBOT-PathPre-training domain=Histopathology, Architecture type=Transformer, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 64 | |
| MoCoV2-ResNet50Pre-training domain=Histopathology, Architecture type=CNN-based, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 62 | |
| Barlow-Twins-ResNet50Pre-training domain=Histopathology, Architecture type=CNN-based, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 58 | |
| PLIPPre-training domain=Histopathology, Architecture type=Transformer, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 58 | |
| DinoSSLPathology-8Pre-training domain=Histopathology, Architecture type=Transformer, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 58 | |
| ConvNext-b-224Pre-training domain=Natural, Architecture type=CNN-based, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 56 | |
| ConvNext-xlargePre-training domain=Natural, Architecture type=CNN-based, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 56 | |
| DinoV1-ViT-b16Pre-training domain=Natural, Architecture type=Transformer, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 55 | |
| DinoV2-ViT-b14Pre-training domain=Natural, Architecture type=Transformer, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 53 | |
| PathDino-224Pre-training domain=Histopathology, Architecture type=Transformer, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 53 | |
| EfficientNet-b5Pre-training domain=Natural, Architecture type=CNN-based, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 51 | |
| MuDiPath-DenseNet-101Pre-training domain=Histopathology, Architecture type=CNN-based, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 51 | |
| KimiaNetPre-training domain=Histopathology, Architecture type=CNN-based, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 51 | |
| CLIP - ViT-B/16Pre-training domain=Natural, Architecture type=Transformer, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 49 | |
| ResNet50Pre-training domain=Natural, Architecture type=CNN-based, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 48 | |
| DenseNet121Pre-training domain=Natural, Architecture type=CNN-based, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 48 | |
| DinoV1-ViT-s16Pre-training domain=Natural, Architecture type=Transformer, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 48 | |
| BiomedCLIPPre-training domain=Histopathology, Architecture type=Transformer, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 47 | |
| MuDiPath-ResNet50Pre-training domain=Histopathology, Architecture type=CNN-based, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 44 | |
| HIPT-ViT-s16Pre-training domain=Histopathology, Architecture type=Transformer, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 44 | |
| EfficientNet-b3-288Pre-training domain=Natural, Architecture type=CNN-based, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 41 | |
| ViT-b16-224Pre-training domain=Natural, Architecture type=Transformer, Patching method=FPS, Aggregation strategy=min-of-medium (Yottixel)2023.11 | 41 |