Submission ID: 5lyipbit
Example: Upright SIFT (OpenCV), 2k features
Processed: 2021-05-11. Download link: 5lyipbit4aqqwju4tusuyd-rootsift-upright-2k-both-degensac.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.
- Phototourism dataset: Stereo track / Multiview track
- Prague Parks dataset: Stereo track / Multiview track
- Google Urban dataset: Stereo track / Multiview track
Metadata
- Authors: (contact)
- Keypoint: sift-def
- Descriptor: rootsift-upright (128 float32: 512 bytes)
- Number of features: 2048
- Summary: SIFT with 2048 features, using the built-in matcher (bidirectional filter with the both strategy, optimal inlier and ratio test thresholds) with DEGENSAC, and setting keypoint orientation to a constant value to increase performance.
- Paper: N/A
- Website: https://opencv.org
- Processing date: 2021-05-11
Phototourism dataset / Stereo track
mAA at 10 degrees: 0.38279 (±0.00000 over 1 run(s) / ±0.12411 over 9 scenes)
Rank (per category): 71 (of 87)
| Scene | Features | Matches (raw) |
Matches (final) |
Rep. @ 3 px. | MS @ 3 px. | mAA(5o) | mAA(10o) |
| BM | 2004.2 | 202.4 | 145.6 | 0.427 Rank: 67/87 |
0.963 Rank: 2/87 |
0.16450 (±0.00000) Rank: 61/87 |
0.27578 (±0.00000) Rank: 60/87 |
| FCS | 2043.3 | 158.8 | 117.5 | 0.305 Rank: 84/87 |
0.856 Rank: 54/87 |
0.45423 (±0.00000) Rank: 67/87 |
0.57527 (±0.00000) Rank: 69/87 |
| LMS | 1027.7 | 106.6 | 71.1 | 0.360 Rank: 68/87 |
0.700 Rank: 13/87 |
0.38010 (±0.00000) Rank: 72/87 |
0.50126 (±0.00000) Rank: 71/87 |
| LB | 1944.7 | 126.6 | 75.2 | 0.311 Rank: 83/87 |
0.686 Rank: 57/87 |
0.29915 (±0.00000) Rank: 61/87 |
0.38323 (±0.00000) Rank: 71/87 |
| MC | 2026.8 | 127.5 | 86.5 | 0.333 Rank: 84/87 |
0.940 Rank: 4/87 |
0.26780 (±0.00000) Rank: 61/87 |
0.40111 (±0.00000) Rank: 63/87 |
| MR | 1950.2 | 186.6 | 144.5 | 0.328 Rank: 84/87 |
0.956 Rank: 1/87 |
0.22559 (±0.00000) Rank: 52/87 |
0.32776 (±0.00000) Rank: 49/87 |
| PSM | 2024.3 | 59.9 | 34.4 | 0.273 Rank: 83/87 |
0.404 Rank: 75/87 |
0.06754 (±0.00000) Rank: 75/87 |
0.12408 (±0.00000) Rank: 77/87 |
| SF | 1992.0 | 199.9 | 137.4 | 0.336 Rank: 77/87 |
0.799 Rank: 24/87 |
0.33649 (±0.00000) Rank: 69/87 |
0.45718 (±0.00000) Rank: 70/87 |
| SPC | 2022.4 | 123.6 | 75.4 | 0.318 Rank: 83/87 |
0.789 Rank: 60/87 |
0.28137 (±0.00000) Rank: 71/87 |
0.39945 (±0.00000) Rank: 73/87 |
| Avg | 1892.8 | 143.5 | 98.6 | 0.333 Rank: 85/87 |
0.788 Rank: 37/87 |
0.27520 (±0.00000) Rank: 70/87 |
0.38279 (±0.00000) Rank: 71/87 |
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 dataset / Multiview track
mAA at 10 degrees: N/A (±0.00000 over 1 run(s) / ±0.17710 over 9 scenes)
Rank (per category): 75 (of 87)
| Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
| BM | 2004.2 | 201.26 | 99.63 Rank: 64/87 |
1025.92 Rank: 70/87 |
5.104 Rank: 22/87 |
0.67582 Rank: 72/87 |
0.29930 (±0.00000) Rank: 74/87 |
0.43911 (±0.00000) Rank: 72/87 |
| FCS | 2043.3 | 167.77 | 93.70 Rank: 77/87 |
1445.39 Rank: 75/87 |
4.124 Rank: 73/87 |
0.32361 Rank: 75/87 |
0.62271 (±0.00000) Rank: 73/87 |
0.69866 (±0.00000) Rank: 74/87 |
| LMS | 1027.7 | 117.21 | 98.03 Rank: 73/87 |
826.36 Rank: 80/87 |
4.556 Rank: 76/87 |
0.31819 Rank: 59/87 |
0.71606 (±0.00000) Rank: 73/87 |
0.80547 (±0.00000) Rank: 72/87 |
| LB | 1944.7 | 150.61 | 94.34 Rank: 76/87 |
1087.40 Rank: 67/87 |
4.124 Rank: 79/87 |
0.54211 Rank: 75/87 |
0.45984 (±0.00000) Rank: 77/87 |
0.56953 (±0.00000) Rank: 78/87 |
| MC | 2026.8 | 123.68 | 95.67 Rank: 75/87 |
998.37 Rank: 76/87 |
4.026 Rank: 78/87 |
0.49925 Rank: 74/87 |
0.38259 (±0.00000) Rank: 74/87 |
0.51527 (±0.00000) Rank: 75/87 |
| MR | 1950.2 | 186.54 | 90.30 Rank: 71/87 |
1608.89 Rank: 54/87 |
3.925 Rank: 78/87 |
0.53786 Rank: 47/87 |
0.30362 (±0.00000) Rank: 73/87 |
0.41438 (±0.00000) Rank: 72/87 |
| PSM | 2024.3 | 57.95 | 66.28 Rank: 80/87 |
684.44 Rank: 78/87 |
2.843 Rank: 76/87 |
0.85472 Rank: 76/87 |
0.14014 (±0.00000) Rank: 77/87 |
0.19077 (±0.00000) Rank: 77/87 |
| SF | 1992.0 | 200.51 | 93.65 Rank: 74/87 |
1660.27 Rank: 71/87 |
4.269 Rank: 76/87 |
0.37462 Rank: 75/87 |
0.60944 (±0.00000) Rank: 75/87 |
0.69708 (±0.00000) Rank: 74/87 |
| SPC | 2022.4 | 126.24 | 94.78 Rank: 72/87 |
1154.10 Rank: 75/87 |
4.087 Rank: 76/87 |
0.59196 Rank: 69/87 |
0.56498 (±0.00000) Rank: 72/87 |
0.66024 (±0.00000) Rank: 72/87 |
| Avg | 1892.8 | 147.98 | 91.82 Rank: 76/87 |
1165.68 Rank: 76/87 |
4.118 Rank: 77/87 |
0.52424 Rank: 72/87 |
0.45541 (±0.00000) Rank: 73/87 |
0.55450 (±0.00000) Rank: 75/87 |
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 10 images. 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 —
Prague Parks dataset / Stereo track
mAA at 10 degrees: 0.41367 (±0.00000 over 1 run(s) / ±0.02890 over 3 scenes)
Rank (per category): 65 (of 87)
| Scene | Features | Matches (raw) |
Matches (final) |
Rep. @ 3 px. | MS @ 3 px. | mAA(5o) | mAA(10o) |
| Lizard | 2048.0 | 155.9 | 122.6 | 0.056 Rank: 4/87 |
0.006 Rank: 66/87 |
0.28100 (±0.00000) Rank: 65/87 |
0.37282 (±0.00000) Rank: 66/87 |
| Pond | 2048.0 | 207.2 | 148.4 | 0.076 Rank: 5/87 |
0.057 Rank: 8/87 |
0.32224 (±0.00000) Rank: 65/87 |
0.43507 (±0.00000) Rank: 64/87 |
| Tree | 2048.0 | 135.1 | 88.3 | 0.049 Rank: 9/87 |
0.018 Rank: 67/87 |
0.36424 (±0.00000) Rank: 63/87 |
0.43311 (±0.00000) Rank: 63/87 |
| Avg | 2048.0 | 166.0 | 119.8 | 0.060 Rank: 4/87 |
0.027 Rank: 47/87 |
0.32249 (±0.00000) Rank: 64/87 |
0.41367 (±0.00000) Rank: 65/87 |
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.
— Lizard —
— Pond —
— Tree —
Prague Parks dataset / Multiview track
mAA at 10 degrees: N/A (±0.00000 over 1 run(s) / ±0.10175 over 3 scenes)
Rank (per category): 67 (of 87)
| Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
| Lizard | 2048.0 | 127.12 | 74.42 Rank: 69/87 |
404.39 Rank: 69/87 |
3.066 Rank: 68/87 |
11.44112 Rank: 2/87 |
0.38182 (±0.00000) Rank: 67/87 |
0.44519 (±0.00000) Rank: 67/87 |
| Pond | 2048.0 | 217.23 | 62.97 Rank: 69/87 |
570.22 Rank: 67/87 |
2.953 Rank: 29/87 |
0.36861 Rank: 2/87 |
0.18491 (±0.00000) Rank: 68/87 |
0.21764 (±0.00000) Rank: 69/87 |
| Tree | 2048.0 | 127.57 | 69.11 Rank: 62/87 |
325.27 Rank: 65/87 |
2.948 Rank: 49/87 |
5.18791 Rank: 3/87 |
0.37740 (±0.00000) Rank: 59/87 |
0.41950 (±0.00000) Rank: 59/87 |
| Avg | 2048.0 | 157.30 | 68.83 Rank: 67/87 |
433.30 Rank: 67/87 |
2.989 Rank: 60/87 |
5.66588 Rank: 2/87 |
0.31471 (±0.00000) Rank: 66/87 |
0.36078 (±0.00000) Rank: 67/87 |
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 10 images. Keypoints are drawn in blue if they are part of the model, and in red otherwise.
— Lizard —
— Pond —
— Tree —
Google Urban dataset / Stereo track
mAA at 10 degrees: 0.25422 (±0.00000 over 1 run(s) / ±0.12013 over 17 scenes)
Rank (per category): 47 (of 87)
| Scene | Features | Matches (raw) |
Matches (final) |
mAA(5o) | mAA(10o) |
| Amsterdam (AMS) | 1723.4 | 151.0 | 94.4 | 0.13214 (±0.00000) Rank: 47/87 |
0.24866 (±0.00000) Rank: 47/87 |
| Bangkok (BGK) | 1582.5 | 144.5 | 84.7 | 0.03772 (±0.00000) Rank: 48/87 |
0.11241 (±0.00000) Rank: 48/87 |
| Barcelona (BCN) | 1644.1 | 124.1 | 77.8 | 0.09817 (±0.00000) Rank: 36/87 |
0.16813 (±0.00000) Rank: 39/87 |
| Buenos Aires (BAR) | 1846.4 | 157.3 | 91.6 | 0.12146 (±0.00000) Rank: 47/87 |
0.22237 (±0.00000) Rank: 48/87 |
| Cambridge (CAM) | 1287.9 | 100.2 | 53.2 | 0.21233 (±0.00000) Rank: 47/87 |
0.32911 (±0.00000) Rank: 46/87 |
| Cannes (CAN) | 1211.8 | 129.1 | 76.8 | 0.30415 (±0.00000) Rank: 44/87 |
0.38894 (±0.00000) Rank: 44/87 |
| Chicago (CHI) | 1399.1 | 60.7 | 23.9 | 0.07377 (±0.00000) Rank: 45/87 |
0.13374 (±0.00000) Rank: 45/87 |
| Helsinki (HEL) | 1702.7 | 203.9 | 125.0 | 0.31443 (±0.00000) Rank: 40/87 |
0.50361 (±0.00000) Rank: 40/87 |
| Madrid (MAD) | 1758.0 | 106.4 | 49.2 | 0.03161 (±0.00000) Rank: 47/87 |
0.08784 (±0.00000) Rank: 46/87 |
| Mountain View (MTV) | 1556.7 | 89.2 | 38.7 | 0.11406 (±0.00000) Rank: 46/87 |
0.18340 (±0.00000) Rank: 47/87 |
| New Orleans (NOR) | 1452.1 | 105.3 | 54.9 | 0.16391 (±0.00000) Rank: 47/87 |
0.21450 (±0.00000) Rank: 47/87 |
| San Francisco (SF) | 1425.6 | 112.2 | 63.9 | 0.07034 (±0.00000) Rank: 49/87 |
0.15696 (±0.00000) Rank: 49/87 |
| Singapore (SG) | 1568.1 | 137.5 | 72.8 | 0.20000 (±0.00000) Rank: 43/87 |
0.30127 (±0.00000) Rank: 42/87 |
| Sydney (SYD) | 1700.0 | 124.3 | 60.6 | 0.18515 (±0.00000) Rank: 46/87 |
0.28559 (±0.00000) Rank: 48/87 |
| Tokyo (TOK) | 1553.8 | 153.4 | 78.9 | 0.37206 (±0.00000) Rank: 46/87 |
0.50980 (±0.00000) Rank: 46/87 |
| Toronto (TOR) | 1705.4 | 97.9 | 39.5 | 0.09920 (±0.00000) Rank: 48/87 |
0.19600 (±0.00000) Rank: 48/87 |
| Zurich (ZRH) | 1694.3 | 146.4 | 92.9 | 0.22337 (±0.00000) Rank: 39/87 |
0.27938 (±0.00000) Rank: 45/87 |
| Average (Avg) | 1577.2 | 126.1 | 69.3 | 0.16199 (±0.00000) Rank: 47/87 |
0.25422 (±0.00000) Rank: 47/87 |
We show the inliers that survive the robust estimation loop (i.e. RANSAC), or those supplied with the submission if using custom matches. As the dataset does not contain depth, we color them based on the symmetric epopilar distance on normalized coordinates: from 0 (green) to 2e-4 (yellow), and in red above it. Note that these occasionally contain false positives.
— Amsterdam —
— Bangkok —
— Barcelona —
— Buenos Aires —
— Cambridge —
— Cannes —
— Chicago —
— Helsinki —
— Madrid —
— Mountain View —
— New Orleans —
— San Francisco —
— Singapore —
— Sydney —
— Tokyo —
— Toronto —
— Zurich —
Google Urban dataset / Multiview track
mAA at 10 degrees: N/A (±0.00000 over 1 run(s) / ±0.08142 over 17 scenes)
Rank (per category): 52 (of 87)
| Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
| AMS | 1723.4 | 66.69 | 45.60 Rank: 52/87 |
210.98 Rank: 52/87 |
2.713 Rank: 52/87 |
6.51971 Rank: 1/87 |
0.00589 (±0.00000) Rank: 52/87 |
0.01871 (±0.00000) Rank: 52/87 |
| BGK | 1582.5 | 67.47 | 79.66 Rank: 52/87 |
620.56 Rank: 52/87 |
3.283 Rank: 52/87 |
9.58920 Rank: 3/87 |
0.01292 (±0.00000) Rank: 50/87 |
0.04749 (±0.00000) Rank: 50/87 |
| BCN | 1644.1 | 54.63 | 65.81 Rank: 50/87 |
365.21 Rank: 52/87 |
3.148 Rank: 50/87 |
21.01468 Rank: 2/87 |
0.04667 (±0.00000) Rank: 50/87 |
0.07687 (±0.00000) Rank: 50/87 |
| BAR | 1846.4 | 66.53 | 46.45 Rank: 52/87 |
223.75 Rank: 52/87 |
2.684 Rank: 52/87 |
5.59593 Rank: 2/87 |
0.00244 (±0.00000) Rank: 51/87 |
0.00669 (±0.00000) Rank: 52/87 |
| CAM | 1287.9 | 41.33 | 44.27 Rank: 52/87 |
151.20 Rank: 52/87 |
2.710 Rank: 52/87 |
5.91850 Rank: 1/87 |
0.00794 (±0.00000) Rank: 52/87 |
0.01490 (±0.00000) Rank: 52/87 |
| CAN | 1211.8 | 56.17 | 45.03 Rank: 52/87 |
179.02 Rank: 52/87 |
2.788 Rank: 52/87 |
4.91147 Rank: 1/87 |
0.00730 (±0.00000) Rank: 52/87 |
0.02153 (±0.00000) Rank: 52/87 |
| CHI | 1399.1 | 29.54 | 39.85 Rank: 52/87 |
86.84 Rank: 52/87 |
2.662 Rank: 52/87 |
21.78100 Rank: 1/87 |
0.00210 (±0.00000) Rank: 50/87 |
0.00496 (±0.00000) Rank: 51/87 |
| HEL | 1702.7 | 111.48 | 78.64 Rank: 50/87 |
919.39 Rank: 52/87 |
3.052 Rank: 52/87 |
9.53525 Rank: 2/87 |
0.18560 (±0.00000) Rank: 46/87 |
0.29761 (±0.00000) Rank: 46/87 |
| MAD | 1758.0 | 44.35 | 56.39 Rank: 52/87 |
299.71 Rank: 52/87 |
2.946 Rank: 52/87 |
2.55454 Rank: 1/87 |
0.00145 (±0.00000) Rank: 52/87 |
0.00522 (±0.00000) Rank: 52/87 |
| MTV | 1556.7 | 42.34 | 40.69 Rank: 52/87 |
177.95 Rank: 52/87 |
2.462 Rank: 52/87 |
3.19390 Rank: 1/87 |
0.00273 (±0.00000) Rank: 52/87 |
0.00615 (±0.00000) Rank: 52/87 |
| NOR | 1452.1 | 48.47 | 39.71 Rank: 52/87 |
127.09 Rank: 52/87 |
2.701 Rank: 52/87 |
3.50905 Rank: 1/87 |
0.00580 (±0.00000) Rank: 52/87 |
0.01176 (±0.00000) Rank: 52/87 |
| SF | 1425.6 | 48.31 | 37.82 Rank: 52/87 |
135.15 Rank: 52/87 |
2.557 Rank: 52/87 |
1.25739 Rank: 1/87 |
0.00100 (±0.00000) Rank: 52/87 |
0.00320 (±0.00000) Rank: 52/87 |
| SG | 1568.1 | 63.14 | 50.34 Rank: 50/87 |
242.38 Rank: 52/87 |
2.840 Rank: 52/87 |
8.82191 Rank: 3/87 |
0.01773 (±0.00000) Rank: 50/87 |
0.03319 (±0.00000) Rank: 50/87 |
| SYD | 1700.0 | 54.39 | 54.82 Rank: 50/87 |
218.99 Rank: 52/87 |
2.748 Rank: 52/87 |
6.58494 Rank: 2/87 |
0.02995 (±0.00000) Rank: 51/87 |
0.05794 (±0.00000) Rank: 51/87 |
| TOK | 1553.8 | 72.15 | 64.88 Rank: 50/87 |
469.30 Rank: 52/87 |
3.057 Rank: 52/87 |
5.61269 Rank: 1/87 |
0.19221 (±0.00000) Rank: 50/87 |
0.23473 (±0.00000) Rank: 50/87 |
| TOR | 1705.4 | 38.72 | 53.36 Rank: 50/87 |
141.76 Rank: 52/87 |
3.129 Rank: 42/87 |
5.06236 Rank: 1/87 |
0.00750 (±0.00000) Rank: 50/87 |
0.01485 (±0.00000) Rank: 51/87 |
| ZRH | 1694.3 | 64.81 | 55.03 Rank: 50/87 |
376.90 Rank: 52/87 |
2.833 Rank: 52/87 |
8.14179 Rank: 1/87 |
0.01381 (±0.00000) Rank: 49/87 |
0.02722 (±0.00000) Rank: 49/87 |
| Avg | 1577.2 | 57.09 | 52.84 Rank: 52/87 |
290.95 Rank: 52/87 |
2.842 Rank: 52/87 |
7.62378 Rank: 1/87 |
0.03194 (±0.00000) Rank: 52/87 |
0.05194 (±0.00000) Rank: 52/87 |
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 10 images. Keypoints are drawn in blue if they are part of the model, and in red otherwise.
— Amsterdam —
— Bangkok —
— Barcelona —
— Buenos Aires —
— Cambridge —
— Cannes —
— Chicago —
— Helsinki —
— Madrid —
— Mountain View —
— New Orleans —
— San Francisco —
— Singapore —
— Sydney —
— Tokyo —
— Toronto —
— Zurich —
