Submission ID: 00022
L2-Net (upright), DEGENSAC
Processed: 20-04-23. Download link: sid-00022-l2net-upright-degensac-2k.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-def
- Descriptor: l2net-upright (128 float32: 512 bytes)
- Number of features: 2048
- Summary: L2-Net descriptors extracted on SIFT keypoints with a fixed orientation, and DEGENSAC for stereo. Re-using parameters found for the orientation-sensitive variant. Please refer to the baselines repository (linked) for details.
- Paper: http://www.nlpr.ia.ac.cn/fanbin/pub/L2-Net_CVPR17.pdf
- Website: https://github.com/vcg-uvic/image-matching-benchmark-baselines
- Origin: Baseline
- Flags: is_baseline
Phototourism / Stereo track
mAA at 10 degrees: 0.40729 (±0.00097 over 3 run(s) / ±0.13173 over 9 scenes)
Rank (per category): 50 (of 108)
Scene | Features | Matches (matcher) |
Matches (filter) |
Matches (final) |
Rep. @ 3 px. | MS @ 3 px. | mAA(5o) | mAA(10o) |
bm | 2004.2 | 222.2 | 222.2 | 155.0 | 0.427 Rank: 56/108 |
0.952 Rank: 9/108 |
0.16427 (±0.00034) Rank: 44/108 |
0.27881 (±0.00241) Rank: 44/108 |
fcs | 2043.3 | 203.0 | 203.0 | 143.4 | 0.305 Rank: 48/108 |
0.884 Rank: 33/108 |
0.50076 (±0.00019) Rank: 45/108 |
0.62450 (±0.00063) Rank: 49/108 |
lms | 1027.6 | 138.1 | 138.1 | 94.8 | 0.360 Rank: 65/108 |
0.706 Rank: 21/108 |
0.40907 (±0.00113) Rank: 49/108 |
0.53733 (±0.00090) Rank: 47/108 |
lb | 1943.7 | 148.4 | 148.4 | 86.0 | 0.312 Rank: 52/108 |
0.698 Rank: 31/108 |
0.31254 (±0.00034) Rank: 44/108 |
0.39920 (±0.00172) Rank: 52/108 |
mc | 2026.8 | 143.2 | 143.2 | 94.0 | 0.333 Rank: 51/108 |
0.926 Rank: 25/108 |
0.27128 (±0.00034) Rank: 44/108 |
0.39970 (±0.00129) Rank: 50/108 |
mr | 1950.2 | 180.4 | 180.4 | 134.8 | 0.328 Rank: 65/108 |
0.941 Rank: 20/108 |
0.21462 (±0.00128) Rank: 45/108 |
0.31147 (±0.00143) Rank: 45/108 |
psm | 2024.3 | 74.7 | 74.7 | 43.3 | 0.273 Rank: 73/108 |
0.430 Rank: 48/108 |
0.08799 (±0.00023) Rank: 61/108 |
0.16821 (±0.00017) Rank: 61/108 |
sf | 1992.0 | 226.9 | 226.9 | 152.1 | 0.336 Rank: 55/108 |
0.825 Rank: 26/108 |
0.36147 (±0.00389) Rank: 49/108 |
0.48759 (±0.00357) Rank: 54/108 |
spc | 2022.4 | 160.6 | 160.6 | 93.2 | 0.318 Rank: 61/108 |
0.845 Rank: 24/108 |
0.32291 (±0.00205) Rank: 53/108 |
0.45884 (±0.00212) Rank: 58/108 |
avg | 1892.7 | 166.4 | 166.4 | 110.7 | 0.333 Rank: 60/108 |
0.801 Rank: 29/108 |
0.29388 (±0.00068) Rank: 49/108 |
0.40729 (±0.00097) Rank: 50/108 |
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: 0.58762 (±0.00228 over 3 run(s) / ±0.17804 over 9 scenes)
Rank (per category): 62 (of 108)
Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
bm | 2004.2 | 220.31 | 99.55 Rank: 66/108 |
1068.48 Rank: 78/108 |
5.141 Rank: 40/108 |
0.65842 Rank: 63/108 |
0.28608 (±0.00818) Rank: 70/108 |
0.42730 (±0.00945) Rank: 71/108 |
fcs | 2043.3 | 213.85 | 95.99 Rank: 72/108 |
1716.01 Rank: 59/108 |
4.227 Rank: 58/108 |
0.28216 Rank: 58/108 |
0.66483 (±0.00351) Rank: 63/108 |
0.74436 (±0.00197) Rank: 61/108 |
lms | 1027.6 | 149.73 | 98.04 Rank: 62/108 |
981.94 Rank: 84/108 |
4.783 Rank: 51/108 |
0.29878 Rank: 31/108 |
0.74968 (±0.00105) Rank: 53/108 |
0.83036 (±0.00376) Rank: 52/108 |
lb | 1943.7 | 180.39 | 96.51 Rank: 72/108 |
1232.41 Rank: 68/108 |
4.277 Rank: 62/108 |
0.52588 Rank: 67/108 |
0.48570 (±0.00594) Rank: 65/108 |
0.60670 (±0.00505) Rank: 66/108 |
mc | 2026.8 | 138.45 | 96.62 Rank: 64/108 |
1138.73 Rank: 69/108 |
4.086 Rank: 68/108 |
0.50998 Rank: 70/108 |
0.39295 (±0.00925) Rank: 64/108 |
0.52626 (±0.01008) Rank: 66/108 |
mr | 1950.2 | 180.96 | 91.96 Rank: 55/108 |
1667.57 Rank: 44/108 |
3.989 Rank: 68/108 |
0.54463 Rank: 36/108 |
0.33064 (±0.00458) Rank: 60/108 |
0.43940 (±0.00380) Rank: 63/108 |
psm | 2024.3 | 72.55 | 76.16 Rank: 69/108 |
980.84 Rank: 75/108 |
2.932 Rank: 64/108 |
0.75429 Rank: 66/108 |
0.18496 (±0.00275) Rank: 69/108 |
0.25633 (±0.00163) Rank: 66/108 |
sf | 1992.0 | 226.42 | 94.78 Rank: 72/108 |
1809.56 Rank: 55/108 |
4.378 Rank: 68/108 |
0.36974 Rank: 74/108 |
0.62686 (±0.00550) Rank: 66/108 |
0.71667 (±0.00550) Rank: 69/108 |
spc | 2022.4 | 164.74 | 98.58 Rank: 54/108 |
1406.53 Rank: 70/108 |
4.219 Rank: 68/108 |
0.53042 Rank: 47/108 |
0.62840 (±0.00525) Rank: 52/108 |
0.74122 (±0.00603) Rank: 51/108 |
avg | 1892.7 | 171.93 | 94.24 Rank: 64/108 |
1333.56 Rank: 68/108 |
4.226 Rank: 67/108 |
0.49715 Rank: 58/108 |
0.48334 (±0.00219) Rank: 62/108 |
0.58762 (±0.00228) Rank: 62/108 |
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 —