Submission ID: 00132
CV-DoG-AffNet-HardNet-kornia-MAGSAC
Processed: 21-02-05. Download link: sid-00132-dog-affnethardnet-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: sift8k
- Descriptor: affnethardnet (128 float32: 512 bytes)
- Number of features: 8000
- Summary: OpenCV DoG keypoints, followed by AffNet shape estimation and HardNet descriptor. Implementation: OpenCV + kornia library
- Paper: https://arxiv.org/abs/1711.06704
- Website: https://kornia.readthedocs.io/en/latest/feature.html
- Origin: Baseline
- Flags: is_baseline
Phototourism / Stereo track
mAA at 10 degrees: 0.54116 (±0.00028 over 3 run(s) / ±0.14669 over 9 scenes)
Rank (per category): 40 (of 147)
| Scene | Features | Matches (matcher) |
Matches (filter) |
Matches (final) |
Rep. @ 3 px. | MS @ 3 px. | mAA(5o) | mAA(10o) |
| bm | 7962.4 | 486.1 | 486.1 | 442.0 | 0.520 Rank: 124/147 |
0.945 Rank: 49/147 |
0.21074 (±0.00559) Rank: 62/147 |
0.33230 (±0.00539) Rank: 59/147 |
| fcs | 7839.9 | 622.6 | 622.6 | 577.0 | 0.465 Rank: 69/147 |
0.937 Rank: 47/147 |
0.61999 (±0.00318) Rank: 49/147 |
0.73316 (±0.00203) Rank: 60/147 |
| lms | 7722.3 | 402.6 | 402.6 | 346.2 | 0.464 Rank: 119/147 |
0.682 Rank: 85/147 |
0.54156 (±0.00235) Rank: 52/147 |
0.65289 (±0.00246) Rank: 57/147 |
| lb | 7938.5 | 412.4 | 412.4 | 354.8 | 0.492 Rank: 69/147 |
0.792 Rank: 23/147 |
0.52286 (±0.00559) Rank: 62/147 |
0.62445 (±0.00098) Rank: 21/147 |
| mc | 7790.2 | 673.7 | 673.7 | 631.9 | 0.550 Rank: 15/147 |
0.963 Rank: 34/147 |
0.41865 (±0.00559) Rank: 62/147 |
0.57157 (±0.00179) Rank: 26/147 |
| mr | 7765.5 | 749.0 | 749.0 | 716.5 | 0.479 Rank: 19/147 |
0.969 Rank: 55/147 |
0.34472 (±0.00205) Rank: 28/147 |
0.46132 (±0.00113) Rank: 26/147 |
| psm | 7771.1 | 358.7 | 358.7 | 327.6 | 0.407 Rank: 19/147 |
0.849 Rank: 3/147 |
0.16247 (±0.00038) Rank: 44/147 |
0.27659 (±0.00250) Rank: 53/147 |
| sf | 7865.6 | 928.5 | 928.5 | 876.0 | 0.499 Rank: 17/147 |
0.891 Rank: 28/147 |
0.54596 (±0.00174) Rank: 16/147 |
0.66919 (±0.00230) Rank: 21/147 |
| spc | 7850.3 | 429.8 | 429.8 | 379.7 | 0.502 Rank: 21/147 |
0.805 Rank: 116/147 |
0.41014 (±0.00173) Rank: 78/147 |
0.54901 (±0.00195) Rank: 85/147 |
| avg | 7834.0 | 562.6 | 562.6 | 516.8 | 0.486 Rank: 74/147 |
0.870 Rank: 19/147 |
0.41968 (±0.00053) Rank: 37/147 |
0.54116 (±0.00028) Rank: 40/147 |
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: 0.72668 (±0.00115 over 3 run(s) / ±0.10796 over 9 scenes)
Rank (per category): 30 (of 147)
| Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
| bm | 7962.4 | 482.17 | 99.71 Rank: 76/147 |
3158.26 Rank: 62/147 |
4.955 Rank: 73/147 |
0.44332 Rank: 40/147 |
0.49437 (±0.00466) Rank: 36/147 |
0.64080 (±0.00920) Rank: 37/147 |
| fcs | 7839.9 | 660.76 | 98.37 Rank: 61/147 |
5280.11 Rank: 52/147 |
4.614 Rank: 46/147 |
0.25022 Rank: 54/147 |
0.74699 (±0.00489) Rank: 19/147 |
0.81172 (±0.00391) Rank: 18/147 |
| lms | 7722.3 | 441.50 | 99.47 Rank: 18/147 |
3693.84 Rank: 48/147 |
4.572 Rank: 35/147 |
0.26096 Rank: 9/147 |
0.85268 (±0.00647) Rank: 6/147 |
0.90690 (±0.00687) Rank: 5/147 |
| lb | 7938.5 | 516.60 | 98.61 Rank: 28/147 |
3624.41 Rank: 36/147 |
4.826 Rank: 29/147 |
0.48983 Rank: 55/147 |
0.69496 (±0.00343) Rank: 35/147 |
0.78784 (±0.00431) Rank: 34/147 |
| mc | 7790.2 | 664.56 | 99.96 Rank: 18/147 |
4848.43 Rank: 36/147 |
4.913 Rank: 29/147 |
0.37338 Rank: 30/147 |
0.53032 (±0.00120) Rank: 28/147 |
0.67341 (±0.00117) Rank: 30/147 |
| mr | 7765.5 | 757.13 | 94.96 Rank: 30/147 |
5962.96 Rank: 24/147 |
4.381 Rank: 29/147 |
0.48381 Rank: 45/147 |
0.46276 (±0.00861) Rank: 32/147 |
0.57800 (±0.01089) Rank: 31/147 |
| psm | 7771.1 | 354.58 | 98.34 Rank: 9/147 |
3944.86 Rank: 45/147 |
3.814 Rank: 3/147 |
0.50805 Rank: 12/147 |
0.50490 (±0.00582) Rank: 23/147 |
0.58884 (±0.00670) Rank: 20/147 |
| sf | 7865.6 | 917.76 | 98.83 Rank: 29/147 |
7252.75 Rank: 22/147 |
4.658 Rank: 26/147 |
0.30324 Rank: 19/147 |
0.74472 (±0.00686) Rank: 21/147 |
0.82735 (±0.00725) Rank: 19/147 |
| spc | 7850.3 | 429.13 | 94.31 Rank: 139/147 |
4276.50 Rank: 59/147 |
4.349 Rank: 60/147 |
0.51417 Rank: 96/147 |
0.63468 (±0.00073) Rank: 108/147 |
0.72526 (±0.00070) Rank: 117/147 |
| avg | 7834.0 | 580.47 | 98.06 Rank: 39/147 |
4671.35 Rank: 38/147 |
4.565 Rank: 33/147 |
0.40300 Rank: 27/147 |
0.62960 (±0.00083) Rank: 31/147 |
0.72668 (±0.00115) Rank: 30/147 |
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 —
