A^2GC: Asymmetric Aggregation with Geometric Constraints for Locally Aggregated Descriptors

Benchmark Model Rank Results
visual-place-recognition-on-mapillary-valA²GC (DINOv2-B, 588x588)#1Recall@1: 96Recall@5: 97.9Recall@10: 98.6
visual-place-recognition-on-pittsburgh-250k-testA²GC (DINOv2-G, 33280-D)#1Recall@1: 97.3Recall@10: 99.7Recall@5: 99.3
visual-place-recognition-on-sf-xl-test-v2A²GC (DINOv2-B, 8448-D)#2Recall@1: 95.2
visual-place-recognition-on-mapillary-testA²GC (DINOv2-G, 33280-D)#6Recall@1: 80.6Recall@5: 90.9Recall@10: 92.5
visual-place-recognition-on-sf-xl-test-v1A²GC (DINOv2-B, 8448-D)#7Recall@1: 89.5
visual-place-recognition-on-spedA²GC (DINOv2-B)#15Recall@1: 83.86Recall@5: 92.75Recall@10: 95.56