Visual Localization on Cambridge Landmarks King’s College

0.19MPE (m)original images

Evaluation results

Method
original images2026.04	0.19	0.3	60.6
Canny2026.04	0.19	0.28	61.2
SAM1 - fine masks2026.04	0.19	0.28	60.9
SAM1 - fine borders2026.04	0.19	0.27	62.1
easy-anon - inpaint2026.04	0.19	0.3	60.9
Canny2026.04	0.19	0.29	60.9
SAM1 - fine borders2026.04	0.19	0.27	61.8
easy-anon - single2026.04	0.2	0.29	60.6
easy-anon - inpaint2026.04	0.2	0.31	60.1
SAM2 - fine borders2026.04	0.2	0.3	58.6
SAM1 - coarse borders2026.04	0.2	0.3	58.6
original images2026.04	0.2	0.3	61.2
blur - 41 px2026.04	0.2	0.3	57.1
easy-anon - single2026.04	0.2	0.3	59.8
SAM2 - fine borders2026.04	0.2	0.3	58
SAM1 - coarse borders2026.04	0.2	0.3	58.3
blur - 41 px2026.04	0.21	0.31	56.3
SAM2 - fine masks2026.04	0.21	0.31	59.2
SAM1 - coarse masks2026.04	0.21	0.32	57.1
SAM1 - fine masks2026.04	0.21	0.29	60.9
Mask2Former - borders2026.04	0.22	0.33	55.7
SAM2 - coarse borders2026.04	0.22	0.33	53.9
SAM2 - fine masks2026.04	0.22	0.31	58.9
SAM1 - coarse masks2026.04	0.22	0.32	57.1
SAM2 - coarse borders2026.04	0.23	0.33	55.1
Mask2Former - semantic2026.04	0.23	0.35	53.9
Mask2Former - semantic2026.04	0.23	0.34	53.9
Mask2Former - borders2026.04	0.23	0.33	54.8
SAM2 - coarse masks2026.04	0.24	0.38	53.1
SAM2 - coarse masks2026.04	0.24	0.36	51.9
Metric3D → Canny2026.04	0.25	0.37	50.1
Metric3D → Canny2026.04	0.25	0.37	51.9
DiffusionEdge2026.04	0.26	0.4	46.9
DiffusionEdge2026.04	0.26	0.41	47.8
Mask2Former - random2026.04	0.28	0.4	44
Mask2Former - random2026.04	0.29	0.4	43.7
pixelization - 10x2026.04	0.37	0.63	31.5
pixelization - 10x2026.04	0.4	0.65	30.6
blur - 81 px2026.04	0.43	0.76	22.2
blur - 81 px2026.04	0.43	0.76	23.6
pixelization - 20x2026.04	1.87	3.08	1.5
pixelization - 20x2026.04	1.94	3.12	2