Cross-image patch-level matching on GenAI
83.49APMultiplex Transformer
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Multiplex TransformerMatching Type=Cross-img, Backbone=ViT-B/16, Pre-trained=ImageNet [3], Number of Layers=4, Patch Size=16x162025.09 | 83.49 | 73.39 | 73.73 | 73.05 | 67.73 | 77.42 |