Submission ID: 00143
CV-DoG-MKD-Concat-magsac
Processed: 21-02-05. Download link: sid-00143-cv-dog-mkd-concat-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: sift
- Descriptor: mkd-concat (128 float32: 512 bytes)
- Number of features: 8000
- Summary: OpenCV DoG keypoints, followed by the MKD-Concat descriptor. Implementation: OpenCV + kornia library
- Paper: https://arxiv.org/abs/1811.11147
- Website: N/A
- Origin: Baseline
- Flags: is_baseline
Phototourism / Stereo track
mAA at 10 degrees: 0.48096 (±0.00051 over 3 run(s) / ±0.15086 over 9 scenes)
Rank (per category): 96 (of 147)
| Scene | Features | Matches (matcher) |
Matches (filter) |
Matches (final) |
Rep. @ 3 px. | MS @ 3 px. | mAA(5o) | mAA(10o) |
| bm | 7968.3 | 462.1 | 462.1 | 414.5 | 0.527 Rank: 40/147 |
0.934 Rank: 82/147 |
0.15078 (±0.00107) Rank: 109/147 |
0.25518 (±0.00153) Rank: 108/147 |
| fcs | 7873.7 | 556.2 | 556.2 | 502.6 | 0.443 Rank: 112/147 |
0.940 Rank: 33/147 |
0.60250 (±0.00122) Rank: 77/147 |
0.72619 (±0.00115) Rank: 73/147 |
| lms | 7770.9 | 356.6 | 356.6 | 301.0 | 0.468 Rank: 44/147 |
0.704 Rank: 28/147 |
0.45091 (±0.00224) Rank: 112/147 |
0.54539 (±0.00287) Rank: 114/147 |
| lb | 7952.5 | 305.2 | 305.2 | 255.4 | 0.474 Rank: 113/147 |
0.739 Rank: 93/147 |
0.36631 (±0.00107) Rank: 109/147 |
0.45000 (±0.00049) Rank: 98/147 |
| mc | 7823.8 | 476.6 | 476.6 | 433.0 | 0.524 Rank: 90/147 |
0.955 Rank: 61/147 |
0.38416 (±0.00107) Rank: 109/147 |
0.53601 (±0.00159) Rank: 63/147 |
| mr | 7788.2 | 487.8 | 487.8 | 469.5 | 0.457 Rank: 84/147 |
0.972 Rank: 46/147 |
0.32303 (±0.00167) Rank: 54/147 |
0.43581 (±0.00217) Rank: 49/147 |
| psm | 7807.2 | 208.7 | 208.7 | 177.3 | 0.385 Rank: 93/147 |
0.531 Rank: 102/147 |
0.12934 (±0.00155) Rank: 94/147 |
0.23031 (±0.00234) Rank: 96/147 |
| sf | 7893.5 | 587.1 | 587.1 | 548.5 | 0.480 Rank: 93/147 |
0.874 Rank: 70/147 |
0.48384 (±0.00304) Rank: 78/147 |
0.61405 (±0.00045) Rank: 84/147 |
| spc | 7868.9 | 387.5 | 387.5 | 330.4 | 0.488 Rank: 81/147 |
0.850 Rank: 83/147 |
0.38773 (±0.00180) Rank: 88/147 |
0.53570 (±0.00112) Rank: 87/147 |
| avg | 7860.8 | 425.3 | 425.3 | 381.4 | 0.472 Rank: 82/147 |
0.833 Rank: 82/147 |
0.36429 (±0.00007) Rank: 94/147 |
0.48096 (±0.00051) Rank: 96/147 |
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: 0.64763 (±0.00344 over 3 run(s) / ±0.16165 over 9 scenes)
Rank (per category): 108 (of 147)
| Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
| bm | 7968.3 | 364.72 | 99.71 Rank: 76/147 |
2666.90 Rank: 96/147 |
4.894 Rank: 90/147 |
0.59079 Rank: 111/147 |
0.35784 (±0.00247) Rank: 108/147 |
0.51114 (±0.00427) Rank: 106/147 |
| fcs | 7873.7 | 469.85 | 98.49 Rank: 51/147 |
4392.67 Rank: 91/147 |
4.399 Rank: 83/147 |
0.24631 Rank: 37/147 |
0.72953 (±0.00775) Rank: 55/147 |
0.80217 (±0.00809) Rank: 40/147 |
| lms | 7770.9 | 304.28 | 98.70 Rank: 94/147 |
2987.57 Rank: 101/147 |
4.315 Rank: 117/147 |
0.28905 Rank: 66/147 |
0.78425 (±0.01006) Rank: 105/147 |
0.85146 (±0.00801) Rank: 100/147 |
| lb | 7952.5 | 295.57 | 97.50 Rank: 112/147 |
2706.00 Rank: 106/147 |
4.348 Rank: 111/147 |
0.53944 Rank: 119/147 |
0.55031 (±0.01040) Rank: 109/147 |
0.64724 (±0.01091) Rank: 108/147 |
| mc | 7823.8 | 367.78 | 99.43 Rank: 96/147 |
3698.89 Rank: 96/147 |
4.352 Rank: 119/147 |
0.43159 Rank: 110/147 |
0.49085 (±0.00324) Rank: 90/147 |
0.64190 (±0.00272) Rank: 87/147 |
| mr | 7788.2 | 417.55 | 94.81 Rank: 42/147 |
4250.17 Rank: 99/147 |
4.002 Rank: 104/147 |
0.46499 Rank: 10/147 |
0.43796 (±0.01244) Rank: 63/147 |
0.55120 (±0.01539) Rank: 60/147 |
| psm | 7807.2 | 137.99 | 80.52 Rank: 126/147 |
2194.49 Rank: 113/147 |
3.012 Rank: 110/147 |
0.80062 Rank: 70/147 |
0.23701 (±0.01317) Rank: 117/147 |
0.30589 (±0.01577) Rank: 121/147 |
| sf | 7893.5 | 479.12 | 96.86 Rank: 116/147 |
5178.35 Rank: 102/147 |
4.124 Rank: 127/147 |
0.37407 Rank: 126/147 |
0.68348 (±0.00626) Rank: 115/147 |
0.77075 (±0.00703) Rank: 115/147 |
| spc | 7868.9 | 295.42 | 98.68 Rank: 94/147 |
3491.47 Rank: 92/147 |
4.076 Rank: 108/147 |
0.50060 Rank: 85/147 |
0.64376 (±0.00307) Rank: 97/147 |
0.74688 (±0.00352) Rank: 100/147 |
| avg | 7860.8 | 348.03 | 96.08 Rank: 120/147 |
3507.39 Rank: 100/147 |
4.169 Rank: 112/147 |
0.47083 Rank: 112/147 |
0.54611 (±0.00334) Rank: 105/147 |
0.64763 (±0.00344) Rank: 108/147 |
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 —
