Submission ID: 5a6d353f
Estelle-2K
Processed: 21-06-11. Download link: 5a6d353f0f5d1dd9-estelle-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.
- Phototourism dataset: Stereo track / Multiview track
- Prague Parks dataset: Stereo track / Multiview track
- Google Urban dataset: Stereo track / Multiview track
Metadata
- Authors: Estelle (contact)
- Keypoint: disk
- Descriptor: disk (128 float32: 512 bytes)
- Number of features: 2048
- Summary: Estelle-2K
- Paper: N/A
- Website: N/A
- Processing date: 21-06-11
Phototourism dataset / Stereo track
mAA at 10 degrees: 0.29313 (±0.00000 over 1 run(s) / ±0.10255 over 9 scenes)
Rank (per category): 79 (of 87)
Scene | Features | Matches (raw) |
Matches (final) |
Rep. @ 3 px. | MS @ 3 px. | mAA(5o) | mAA(10o) |
BM | 2048.0 | — | 522.1 | 0.605 Rank: 6/87 |
0.954 Rank: 4/87 |
0.08366 (±0.00000) Rank: 78/87 |
0.16591 (±0.00000) Rank: 78/87 |
FCS | 2048.0 | — | 281.9 | 0.407 Rank: 34/87 |
0.783 Rank: 74/87 |
0.23165 (±0.00000) Rank: 84/87 |
0.32210 (±0.00000) Rank: 84/87 |
LMS | 2048.0 | — | 234.9 | 0.380 Rank: 60/87 |
0.648 Rank: 70/87 |
0.26689 (±0.00000) Rank: 82/87 |
0.34900 (±0.00000) Rank: 83/87 |
LB | 2045.5 | — | 329.0 | 0.448 Rank: 11/87 |
0.738 Rank: 15/87 |
0.25667 (±0.00000) Rank: 78/87 |
0.34977 (±0.00000) Rank: 74/87 |
MC | 2048.0 | — | 469.4 | 0.543 Rank: 3/87 |
0.912 Rank: 23/87 |
0.22439 (±0.00000) Rank: 78/87 |
0.35281 (±0.00000) Rank: 68/87 |
MR | 2048.0 | — | 298.6 | 0.435 Rank: 18/87 |
0.799 Rank: 77/87 |
0.18329 (±0.00000) Rank: 73/87 |
0.25590 (±0.00000) Rank: 76/87 |
PSM | 2048.0 | — | 163.3 | 0.370 Rank: 13/87 |
0.538 Rank: 57/87 |
0.04220 (±0.00000) Rank: 82/87 |
0.08174 (±0.00000) Rank: 83/87 |
SF | 2046.8 | — | 256.1 | 0.372 Rank: 41/87 |
0.705 Rank: 75/87 |
0.23398 (±0.00000) Rank: 79/87 |
0.32393 (±0.00000) Rank: 80/87 |
SPC | 2041.0 | — | 421.6 | 0.467 Rank: 6/87 |
0.838 Rank: 15/87 |
0.29395 (±0.00000) Rank: 69/87 |
0.43706 (±0.00000) Rank: 69/87 |
Avg | 2046.8 | — | 330.8 | 0.448 Rank: 9/87 |
0.768 Rank: 64/87 |
0.20185 (±0.00000) Rank: 78/87 |
0.29313 (±0.00000) Rank: 79/87 |
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) / ±0.11163 over 9 scenes)
Rank (per category): 67 (of 87)
Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
BM | 2048.0 | 522.80 | 99.73 Rank: 56/87 |
1725.23 Rank: 49/87 |
7.177 Rank: 3/87 |
0.59373 Rank: 63/87 |
0.37168 (±0.00000) Rank: 64/87 |
0.53746 (±0.00000) Rank: 64/87 |
FCS | 2048.0 | 298.08 | 94.95 Rank: 73/87 |
1956.88 Rank: 58/87 |
5.154 Rank: 13/87 |
0.27557 Rank: 62/87 |
0.65629 (±0.00000) Rank: 69/87 |
0.72846 (±0.00000) Rank: 70/87 |
LMS | 2048.0 | 257.18 | 97.18 Rank: 75/87 |
1305.09 Rank: 67/87 |
5.756 Rank: 27/87 |
0.35283 Rank: 74/87 |
0.66163 (±0.00000) Rank: 77/87 |
0.74174 (±0.00000) Rank: 77/87 |
LB | 2045.5 | 400.84 | 96.28 Rank: 70/87 |
1504.48 Rank: 52/87 |
6.331 Rank: 1/87 |
0.47239 Rank: 15/87 |
0.62403 (±0.00000) Rank: 62/87 |
0.72471 (±0.00000) Rank: 63/87 |
MC | 2048.0 | 459.02 | 98.41 Rank: 68/87 |
2015.99 Rank: 44/87 |
6.418 Rank: 2/87 |
0.40785 Rank: 58/87 |
0.47197 (±0.00000) Rank: 63/87 |
0.62236 (±0.00000) Rank: 62/87 |
MR | 2048.0 | 299.09 | 90.41 Rank: 70/87 |
2064.39 Rank: 42/87 |
4.910 Rank: 22/87 |
0.57103 Rank: 65/87 |
0.37889 (±0.00000) Rank: 53/87 |
0.48265 (±0.00000) Rank: 60/87 |
PSM | 2048.0 | 161.66 | 86.88 Rank: 67/87 |
1881.19 Rank: 57/87 |
4.182 Rank: 49/87 |
0.62149 Rank: 65/87 |
0.40387 (±0.00000) Rank: 61/87 |
0.47214 (±0.00000) Rank: 63/87 |
SF | 2046.8 | 250.33 | 92.75 Rank: 76/87 |
1962.18 Rank: 59/87 |
4.942 Rank: 31/87 |
0.35642 Rank: 73/87 |
0.59777 (±0.00000) Rank: 76/87 |
0.68896 (±0.00000) Rank: 76/87 |
SPC | 2041.0 | 438.39 | 98.35 Rank: 65/87 |
2058.08 Rank: 48/87 |
6.000 Rank: 3/87 |
0.43620 Rank: 29/87 |
0.68190 (±0.00000) Rank: 57/87 |
0.78387 (±0.00000) Rank: 61/87 |
Avg | 2046.8 | 343.04 | 94.99 Rank: 69/87 |
1830.39 Rank: 52/87 |
5.652 Rank: 3/87 |
0.45417 Rank: 64/87 |
0.53867 (±0.00000) Rank: 67/87 |
0.64248 (±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.
— 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: 0.21261 (±0.00000 over 1 run(s) / ±0.01401 over 3 scenes)
Rank (per category): 72 (of 87)
Scene | Features | Matches (raw) |
Matches (final) |
Rep. @ 3 px. | MS @ 3 px. | mAA(5o) | mAA(10o) |
Lizard | 2048.0 | — | 109.8 | 0.045 Rank: 7/87 |
0.004 Rank: 71/87 |
0.13592 (±0.00000) Rank: 73/87 |
0.19681 (±0.00000) Rank: 73/87 |
Pond | 2048.0 | — | 123.9 | 0.075 Rank: 6/87 |
0.057 Rank: 6/87 |
0.16072 (±0.00000) Rank: 72/87 |
0.23086 (±0.00000) Rank: 72/87 |
Tree | 2048.0 | — | 61.7 | 0.049 Rank: 4/87 |
0.017 Rank: 69/87 |
0.14967 (±0.00000) Rank: 72/87 |
0.21015 (±0.00000) Rank: 72/87 |
Avg | 2048.0 | — | 98.5 | 0.056 Rank: 6/87 |
0.026 Rank: 57/87 |
0.14877 (±0.00000) Rank: 72/87 |
0.21261 (±0.00000) Rank: 72/87 |
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) / ±0.11824 over 3 scenes)
Rank (per category): 69 (of 87)
Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
Lizard | 2048.0 | 143.34 | 77.84 Rank: 67/87 |
474.27 Rank: 67/87 |
3.240 Rank: 59/87 |
12.93227 Rank: 5/87 |
0.34407 (±0.00000) Rank: 68/87 |
0.43856 (±0.00000) Rank: 68/87 |
Pond | 2048.0 | 234.33 | 65.97 Rank: 66/87 |
719.64 Rank: 62/87 |
3.020 Rank: 16/87 |
0.39670 Rank: 4/87 |
0.19169 (±0.00000) Rank: 67/87 |
0.23483 (±0.00000) Rank: 67/87 |
Tree | 2048.0 | 86.93 | 61.97 Rank: 68/87 |
243.72 Rank: 68/87 |
2.933 Rank: 53/87 |
4.60344 Rank: 2/87 |
0.13829 (±0.00000) Rank: 69/87 |
0.15842 (±0.00000) Rank: 69/87 |
Avg | 2048.0 | 154.87 | 68.59 Rank: 68/87 |
479.21 Rank: 66/87 |
3.064 Rank: 41/87 |
5.97747 Rank: 4/87 |
0.22468 (±0.00000) Rank: 69/87 |
0.27727 (±0.00000) Rank: 69/87 |
Google Urban dataset / Stereo track
mAA at 10 degrees: 0.11209 (±0.00000 over 1 run(s) / ±0.06059 over 17 scenes)
Rank (per category): 52 (of 87)
Scene | Features | Matches (raw) |
Matches (final) |
mAA(5o) | mAA(10o) |
Amsterdam (AMS) | 2048.0 | — | 148.6 | 0.07143 (±0.00000) Rank: 51/87 |
0.14196 (±0.00000) Rank: 51/87 |
Bangkok (BGK) | 2048.0 | — | 168.9 | 0.01439 (±0.00000) Rank: 52/87 |
0.06303 (±0.00000) Rank: 51/87 |
Barcelona (BCN) | 2048.0 | — | 85.0 | 0.02418 (±0.00000) Rank: 52/87 |
0.04908 (±0.00000) Rank: 52/87 |
Buenos Aires (BAR) | 2048.0 | — | 110.4 | 0.06393 (±0.00000) Rank: 51/87 |
0.12466 (±0.00000) Rank: 51/87 |
Cambridge (CAM) | 2045.0 | — | 98.2 | 0.06370 (±0.00000) Rank: 51/87 |
0.12158 (±0.00000) Rank: 52/87 |
Cannes (CAN) | 2048.0 | — | 137.0 | 0.12535 (±0.00000) Rank: 52/87 |
0.21106 (±0.00000) Rank: 52/87 |
Chicago (CHI) | 2042.9 | — | 24.7 | 0.01967 (±0.00000) Rank: 51/87 |
0.04993 (±0.00000) Rank: 51/87 |
Helsinki (HEL) | 2048.0 | — | 212.7 | 0.13763 (±0.00000) Rank: 49/87 |
0.25567 (±0.00000) Rank: 49/87 |
Madrid (MAD) | 2048.0 | — | 56.3 | 0.00729 (±0.00000) Rank: 52/87 |
0.03009 (±0.00000) Rank: 51/87 |
Mountain View (MTV) | 2048.0 | — | 74.0 | 0.04141 (±0.00000) Rank: 51/87 |
0.08027 (±0.00000) Rank: 51/87 |
New Orleans (NOR) | 2048.0 | — | 61.2 | 0.07160 (±0.00000) Rank: 51/87 |
0.10621 (±0.00000) Rank: 52/87 |
San Francisco (SF) | 2048.0 | — | 106.1 | 0.02625 (±0.00000) Rank: 52/87 |
0.06562 (±0.00000) Rank: 52/87 |
Singapore (SG) | 2042.4 | — | 103.8 | 0.06772 (±0.00000) Rank: 51/87 |
0.11171 (±0.00000) Rank: 51/87 |
Sydney (SYD) | 2048.0 | — | 79.7 | 0.04803 (±0.00000) Rank: 51/87 |
0.09258 (±0.00000) Rank: 52/87 |
Tokyo (TOK) | 2048.0 | — | 97.6 | 0.11765 (±0.00000) Rank: 52/87 |
0.19804 (±0.00000) Rank: 52/87 |
Toronto (TOR) | 2048.0 | — | 36.5 | 0.03600 (±0.00000) Rank: 51/87 |
0.06520 (±0.00000) Rank: 51/87 |
Zurich (ZRH) | 2048.0 | — | 115.0 | 0.09141 (±0.00000) Rank: 51/87 |
0.13883 (±0.00000) Rank: 52/87 |
Average (Avg) | 2047.2 | — | 100.9 | 0.06045 (±0.00000) Rank: 52/87 |
0.11209 (±0.00000) Rank: 52/87 |
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) / ±0.06264 over 17 scenes)
Rank (per category): 51 (of 87)
Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
AMS | 2048.0 | 83.10 | 60.22 Rank: 51/87 |
569.97 Rank: 48/87 |
3.405 Rank: 47/87 |
13.63673 Rank: 2/87 |
0.09285 (±0.00000) Rank: 51/87 |
0.13448 (±0.00000) Rank: 51/87 |
BGK | 2048.0 | 105.12 | 86.88 Rank: 50/87 |
1066.26 Rank: 49/87 |
3.930 Rank: 2/87 |
8.16470 Rank: 2/87 |
0.00644 (±0.00000) Rank: 51/87 |
0.02742 (±0.00000) Rank: 51/87 |
BCN | 2048.0 | 45.14 | 59.34 Rank: 51/87 |
446.34 Rank: 49/87 |
3.142 Rank: 51/87 |
17.71361 Rank: 1/87 |
0.03877 (±0.00000) Rank: 52/87 |
0.07352 (±0.00000) Rank: 52/87 |
BAR | 2048.0 | 56.39 | 46.70 Rank: 51/87 |
322.23 Rank: 50/87 |
2.724 Rank: 51/87 |
5.11882 Rank: 1/87 |
0.00230 (±0.00000) Rank: 52/87 |
0.01044 (±0.00000) Rank: 51/87 |
CAM | 2045.0 | 52.74 | 56.77 Rank: 51/87 |
431.05 Rank: 50/87 |
3.281 Rank: 44/87 |
10.49998 Rank: 2/87 |
0.02648 (±0.00000) Rank: 51/87 |
0.04468 (±0.00000) Rank: 51/87 |
CAN | 2048.0 | 67.25 | 55.44 Rank: 47/87 |
451.70 Rank: 38/87 |
3.211 Rank: 42/87 |
9.25697 Rank: 4/87 |
0.00874 (±0.00000) Rank: 51/87 |
0.02998 (±0.00000) Rank: 51/87 |
CHI | 2042.9 | 18.59 | 45.61 Rank: 50/87 |
149.99 Rank: 50/87 |
3.130 Rank: 50/87 |
34.10368 Rank: 2/87 |
0.00107 (±0.00000) Rank: 51/87 |
0.00473 (±0.00000) Rank: 52/87 |
HEL | 2048.0 | 146.59 | 73.02 Rank: 52/87 |
1097.47 Rank: 50/87 |
3.609 Rank: 48/87 |
10.88580 Rank: 6/87 |
0.12035 (±0.00000) Rank: 49/87 |
0.18346 (±0.00000) Rank: 49/87 |
MAD | 2048.0 | 31.10 | 58.68 Rank: 51/87 |
331.51 Rank: 51/87 |
3.294 Rank: 20/87 |
3.05288 Rank: 3/87 |
0.00555 (±0.00000) Rank: 51/87 |
0.01692 (±0.00000) Rank: 51/87 |
MTV | 2048.0 | 51.48 | 69.12 Rank: 49/87 |
576.76 Rank: 47/87 |
3.570 Rank: 46/87 |
10.41085 Rank: 4/87 |
0.05757 (±0.00000) Rank: 50/87 |
0.11653 (±0.00000) Rank: 49/87 |
NOR | 2048.0 | 38.24 | 40.22 Rank: 51/87 |
243.41 Rank: 49/87 |
2.743 Rank: 51/87 |
5.22540 Rank: 2/87 |
0.01420 (±0.00000) Rank: 51/87 |
0.02427 (±0.00000) Rank: 51/87 |
SF | 2048.0 | 60.88 | 49.42 Rank: 51/87 |
296.28 Rank: 51/87 |
3.059 Rank: 50/87 |
5.57780 Rank: 2/87 |
0.01733 (±0.00000) Rank: 51/87 |
0.03407 (±0.00000) Rank: 51/87 |
SG | 2042.4 | 59.68 | 46.21 Rank: 51/87 |
378.64 Rank: 50/87 |
2.986 Rank: 51/87 |
6.27010 Rank: 2/87 |
0.01133 (±0.00000) Rank: 52/87 |
0.02343 (±0.00000) Rank: 52/87 |
SYD | 2048.0 | 40.44 | 52.33 Rank: 52/87 |
248.44 Rank: 51/87 |
2.883 Rank: 51/87 |
6.60989 Rank: 3/87 |
0.04290 (±0.00000) Rank: 50/87 |
0.06798 (±0.00000) Rank: 50/87 |
TOK | 2048.0 | 56.11 | 60.74 Rank: 52/87 |
554.27 Rank: 51/87 |
3.427 Rank: 48/87 |
5.80914 Rank: 2/87 |
0.17046 (±0.00000) Rank: 52/87 |
0.21661 (±0.00000) Rank: 52/87 |
TOR | 2048.0 | 18.71 | 45.13 Rank: 51/87 |
204.31 Rank: 51/87 |
2.796 Rank: 51/87 |
7.93146 Rank: 5/87 |
0.00373 (±0.00000) Rank: 52/87 |
0.01042 (±0.00000) Rank: 52/87 |
ZRH | 2048.0 | 60.34 | 49.99 Rank: 52/87 |
441.23 Rank: 50/87 |
3.116 Rank: 47/87 |
8.27218 Rank: 2/87 |
0.00219 (±0.00000) Rank: 52/87 |
0.00855 (±0.00000) Rank: 52/87 |
Avg | 2047.2 | 58.35 | 56.23 Rank: 51/87 |
459.40 Rank: 51/87 |
3.194 Rank: 49/87 |
9.91412 Rank: 2/87 |
0.03660 (±0.00000) Rank: 51/87 |
0.06044 (±0.00000) Rank: 51/87 |
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 —









