Submission ID: 00118
VL-HessAffNet-SIFT-8k
Processed: 20-05-04. Download link: sid-00118-vl-hessaffnet-sift-8k-both-magsac.json
This page ranks the submission against all others using the same number of keypoints, regardless of descriptor size. Please hover over table headers for descriptions on metrics and full scene names.
Metadata
- Authors: Challenge organizers (contact)
- Keypoint: hessian
- Descriptor: affnetvlsift (128 float32: 512 bytes)
- Number of features: 8000
- Summary: VL-HessAffNet-SIFT with 8000 features, using the built-in matcher (bidirectional filter with the both strategy, optimal inlier and ratio test thresholds) with MAGSAC
- Paper: N/A
- Website: N/A
- Origin: Baseline
- Flags: is_baseline
Phototourism / Stereo track
mAA at 10 degrees: 0.46263 (±0.00079 over 3 run(s) / ±0.14674 over 9 scenes)
Rank (per category): 108 (of 147)
| Scene | Features | Matches (matcher) |
Matches (filter) |
Matches (final) |
Rep. @ 3 px. | MS @ 3 px. | mAA(5o) | mAA(10o) |
| bm | 8000.0 | 534.8 | 534.8 | 443.2 | 0.668 Rank: 7/147 |
0.955 Rank: 13/147 |
0.16032 (±0.00252) Rank: 105/147 |
0.27896 (±0.00371) Rank: 97/147 |
| fcs | 8000.0 | 457.3 | 457.3 | 357.8 | 0.524 Rank: 7/147 |
0.924 Rank: 84/147 |
0.57395 (±0.00278) Rank: 96/147 |
0.69666 (±0.00194) Rank: 96/147 |
| lms | 8000.0 | 570.3 | 570.3 | 418.6 | 0.661 Rank: 1/147 |
0.721 Rank: 7/147 |
0.42998 (±0.00201) Rank: 115/147 |
0.52670 (±0.00176) Rank: 117/147 |
| lb | 8000.0 | 374.5 | 374.5 | 260.4 | 0.555 Rank: 2/147 |
0.737 Rank: 97/147 |
0.36927 (±0.00252) Rank: 105/147 |
0.45370 (±0.00110) Rank: 94/147 |
| mc | 8000.0 | 418.4 | 418.4 | 324.4 | 0.588 Rank: 11/147 |
0.957 Rank: 55/147 |
0.37722 (±0.00252) Rank: 105/147 |
0.52245 (±0.00079) Rank: 81/147 |
| mr | 8000.0 | 456.7 | 456.7 | 407.7 | 0.570 Rank: 10/147 |
0.955 Rank: 86/147 |
0.26950 (±0.00194) Rank: 105/147 |
0.37344 (±0.00258) Rank: 105/147 |
| psm | 8000.0 | 222.5 | 222.5 | 149.8 | 0.494 Rank: 3/147 |
0.423 Rank: 126/147 |
0.11463 (±0.00092) Rank: 111/147 |
0.20161 (±0.00067) Rank: 113/147 |
| sf | 8000.0 | 618.1 | 618.1 | 531.9 | 0.549 Rank: 6/147 |
0.882 Rank: 48/147 |
0.47553 (±0.00054) Rank: 85/147 |
0.60164 (±0.00115) Rank: 91/147 |
| spc | 8000.0 | 357.8 | 357.8 | 256.5 | 0.584 Rank: 3/147 |
0.862 Rank: 61/147 |
0.36525 (±0.00153) Rank: 105/147 |
0.50853 (±0.00247) Rank: 107/147 |
| avg | 8000.0 | 445.6 | 445.6 | 350.0 | 0.577 Rank: 6/147 |
0.824 Rank: 100/147 |
0.34841 (±0.00102) Rank: 105/147 |
0.46263 (±0.00079) Rank: 108/147 |
We show the inliers that survive the robust estimation loop (i.e. RANSAC), or those supplied with the submission if using custom matches, and use the depth estimates to determine whether they are correct. We draw matches above a 5-pixel error threshold in red, and those below are color-coded by their error, from 0 (green) to 5 pixels (yellow). Matches for which we do not have depth estimates are drawn in blue. Please note that the depth maps are estimates and may contain errors.
— british museum —
— florence cathedral side —
— lincoln memorial statue —
— london bridge —
— milan cathedral —
— mount rushmore —
— piazza san marco —
— sagrada familia —
— saint paul's cathedral —
Phototourism / Multiview track
mAA at 10 degrees: 0.60691 (±0.00338 over 3 run(s) / ±0.18216 over 9 scenes)
Rank (per category): 123 (of 147)
| Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
| bm | 8000.0 | 412.93 | 99.38 Rank: 123/147 |
2746.98 Rank: 90/147 |
5.017 Rank: 55/147 |
0.67808 Rank: 127/147 |
0.29625 (±0.01023) Rank: 123/147 |
0.44466 (±0.01458) Rank: 123/147 |
| fcs | 8000.0 | 364.50 | 97.13 Rank: 121/147 |
3769.37 Rank: 113/147 |
4.047 Rank: 126/147 |
0.25664 Rank: 73/147 |
0.69233 (±0.00576) Rank: 111/147 |
0.76599 (±0.00815) Rank: 106/147 |
| lms | 8000.0 | 471.63 | 98.72 Rank: 90/147 |
3803.73 Rank: 39/147 |
4.463 Rank: 68/147 |
0.32260 Rank: 113/147 |
0.75145 (±0.00766) Rank: 121/147 |
0.82202 (±0.00894) Rank: 120/147 |
| lb | 8000.0 | 335.23 | 96.28 Rank: 134/147 |
2791.96 Rank: 97/147 |
4.236 Rank: 126/147 |
0.55347 Rank: 129/147 |
0.51079 (±0.00965) Rank: 121/147 |
0.60827 (±0.01181) Rank: 124/147 |
| mc | 8000.0 | 304.29 | 98.97 Rank: 127/147 |
2905.48 Rank: 123/147 |
4.184 Rank: 129/147 |
0.41768 Rank: 94/147 |
0.46749 (±0.00331) Rank: 115/147 |
0.61109 (±0.00407) Rank: 115/147 |
| mr | 8000.0 | 376.27 | 92.85 Rank: 108/147 |
4211.59 Rank: 105/147 |
3.766 Rank: 135/147 |
0.52127 Rank: 102/147 |
0.34595 (±0.00893) Rank: 123/147 |
0.45492 (±0.01332) Rank: 123/147 |
| psm | 8000.0 | 134.10 | 77.71 Rank: 134/147 |
1894.59 Rank: 129/147 |
2.796 Rank: 132/147 |
0.81506 Rank: 70/147 |
0.16225 (±0.00630) Rank: 134/147 |
0.24306 (±0.00758) Rank: 134/147 |
| sf | 8000.0 | 500.28 | 97.30 Rank: 93/147 |
4839.18 Rank: 114/147 |
4.246 Rank: 106/147 |
0.34601 Rank: 106/147 |
0.68359 (±0.00786) Rank: 112/147 |
0.77516 (±0.00854) Rank: 111/147 |
| spc | 8000.0 | 256.99 | 98.14 Rank: 118/147 |
2986.55 Rank: 117/147 |
3.925 Rank: 128/147 |
0.48885 Rank: 77/147 |
0.63410 (±0.00530) Rank: 114/147 |
0.73706 (±0.00764) Rank: 108/147 |
| avg | 8000.0 | 350.69 | 95.16 Rank: 130/147 |
3327.71 Rank: 111/147 |
4.076 Rank: 124/147 |
0.48885 Rank: 123/147 |
0.50491 (±0.00285) Rank: 122/147 |
0.60691 (±0.00338) Rank: 123/147 |
In the multi-view track we reconstruct the scene with Structure-from-Motion (Colmap) with small sets of images. We show the results for one bag of 25 images (displaying: 10). Keypoints are drawn in blue if they are part of the model, and in red otherwise.
— british museum —
— florence cathedral side —
— lincoln memorial statue —
— london bridge —
— milan cathedral —
— mount rushmore —
— piazza san marco —
— sagrada familia —
— saint paul's cathedral —
