Submission ID: 93ec4ec0
Cassius-8K
Processed: 21-06-16. Download link: 93ec4ec0a1e94ec2-cassius-8k.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: Cassius (contact)
- Keypoint: trial
- Descriptor: trial (128 float32: 512 bytes)
- Number of features: 8000
- Summary: Cassius-8K
- Paper: N/A
- Website: N/A
- Processing date: 21-06-16
Phototourism dataset / Stereo track
mAA at 10 degrees: 0.37448 (±0.00000 over 1 run(s) / ±0.11717 over 9 scenes)
Rank (per category): 41 (of 51)
Scene | Features | Matches (raw) |
Matches (final) |
Rep. @ 3 px. | MS @ 3 px. | mAA(5o) | mAA(10o) |
BM | 6278.6 | — | 1784.8 | 0.682 Rank: 10/51 |
0.925 Rank: 27/51 |
0.14534 (±0.00000) Rank: 40/51 |
0.25795 (±0.00000) Rank: 40/51 |
FCS | 7386.1 | — | 1311.1 | 0.556 Rank: 6/51 |
0.845 Rank: 40/51 |
0.32071 (±0.00000) Rank: 42/51 |
0.42290 (±0.00000) Rank: 42/51 |
LMS | 6403.7 | — | 810.2 | 0.533 Rank: 12/51 |
0.608 Rank: 43/51 |
0.40906 (±0.00000) Rank: 38/51 |
0.51568 (±0.00000) Rank: 38/51 |
LB | 5785.4 | — | 930.2 | 0.586 Rank: 16/51 |
0.713 Rank: 35/51 |
0.26100 (±0.00000) Rank: 40/51 |
0.35982 (±0.00000) Rank: 41/51 |
MC | 6476.0 | — | 1704.5 | 0.691 Rank: 7/51 |
0.911 Rank: 30/51 |
0.25584 (±0.00000) Rank: 40/51 |
0.39242 (±0.00000) Rank: 39/51 |
MR | 7403.5 | — | 1392.4 | 0.640 Rank: 6/51 |
0.858 Rank: 43/51 |
0.22637 (±0.00000) Rank: 39/51 |
0.31721 (±0.00000) Rank: 40/51 |
PSM | 7260.8 | — | 773.2 | 0.538 Rank: 7/51 |
0.646 Rank: 26/51 |
0.07110 (±0.00000) Rank: 40/51 |
0.13924 (±0.00000) Rank: 41/51 |
SF | 7582.5 | — | 1203.9 | 0.559 Rank: 8/51 |
0.749 Rank: 39/51 |
0.32881 (±0.00000) Rank: 40/51 |
0.43171 (±0.00000) Rank: 40/51 |
SPC | 6520.9 | — | 1541.3 | 0.600 Rank: 9/51 |
0.837 Rank: 24/51 |
0.36659 (±0.00000) Rank: 38/51 |
0.53343 (±0.00000) Rank: 35/51 |
Avg | 6788.6 | — | 1272.4 | 0.598 Rank: 7/51 |
0.788 Rank: 38/51 |
0.26498 (±0.00000) Rank: 40/51 |
0.37448 (±0.00000) Rank: 41/51 |
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) / ±0.09858 over 9 scenes)
Rank (per category): 25 (of 51)
Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
BM | 6278.6 | 1798.00 | 99.67 Rank: 34/51 |
5631.00 Rank: 14/51 |
6.711 Rank: 2/51 |
0.44258 Rank: 22/51 |
0.49077 (±0.00000) Rank: 24/51 |
0.63840 (±0.00000) Rank: 24/51 |
FCS | 7386.1 | 1377.04 | 97.83 Rank: 24/51 |
7114.18 Rank: 16/51 |
5.701 Rank: 3/51 |
0.28551 Rank: 35/51 |
0.70822 (±0.00000) Rank: 28/51 |
0.77947 (±0.00000) Rank: 25/51 |
LMS | 6403.7 | 892.56 | 99.73 Rank: 1/51 |
5305.21 Rank: 12/51 |
5.364 Rank: 5/51 |
0.35554 Rank: 39/51 |
0.80085 (±0.00000) Rank: 28/51 |
0.87010 (±0.00000) Rank: 23/51 |
LB | 5785.4 | 1130.53 | 98.03 Rank: 24/51 |
4329.41 Rank: 16/51 |
6.144 Rank: 4/51 |
0.46480 Rank: 7/51 |
0.69809 (±0.00000) Rank: 24/51 |
0.79210 (±0.00000) Rank: 24/51 |
MC | 6476.0 | 1675.88 | 100.00 Rank: 1/51 |
6074.79 Rank: 15/51 |
6.689 Rank: 3/51 |
0.34907 Rank: 19/51 |
0.52827 (±0.00000) Rank: 31/51 |
0.68201 (±0.00000) Rank: 27/51 |
MR | 7403.5 | 1395.03 | 93.52 Rank: 36/51 |
6720.97 Rank: 16/51 |
5.688 Rank: 5/51 |
0.49228 Rank: 24/51 |
0.46166 (±0.00000) Rank: 20/51 |
0.57269 (±0.00000) Rank: 20/51 |
PSM | 7260.8 | 775.35 | 95.49 Rank: 28/51 |
6667.72 Rank: 16/51 |
4.701 Rank: 6/51 |
0.46703 Rank: 19/51 |
0.52535 (±0.00000) Rank: 18/51 |
0.60483 (±0.00000) Rank: 20/51 |
SF | 7582.5 | 1172.16 | 95.84 Rank: 35/51 |
7305.96 Rank: 16/51 |
5.607 Rank: 4/51 |
0.31158 Rank: 27/51 |
0.69822 (±0.00000) Rank: 32/51 |
0.77953 (±0.00000) Rank: 32/51 |
SPC | 6520.9 | 1600.79 | 99.93 Rank: 14/51 |
6607.14 Rank: 15/51 |
6.256 Rank: 2/51 |
0.40028 Rank: 7/51 |
0.71534 (±0.00000) Rank: 21/51 |
0.81990 (±0.00000) Rank: 20/51 |
Avg | 6788.6 | 1313.04 | 97.78 Rank: 31/51 |
6195.15 Rank: 16/51 |
5.873 Rank: 5/51 |
0.39652 Rank: 20/51 |
0.62520 (±0.00000) Rank: 25/51 |
0.72656 (±0.00000) Rank: 25/51 |
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: 0.41780 (±0.00000 over 1 run(s) / ±0.04095 over 3 scenes)
Rank (per category): 35 (of 51)
Scene | Features | Matches (raw) |
Matches (final) |
Rep. @ 3 px. | MS @ 3 px. | mAA(5o) | mAA(10o) |
Lizard | 8000.0 | — | 639.6 | 0.134 Rank: 6/51 |
0.010 Rank: 24/51 |
0.28488 (±0.00000) Rank: 32/51 |
0.37379 (±0.00000) Rank: 34/51 |
Pond | 8000.0 | — | 814.0 | 0.171 Rank: 5/51 |
0.056 Rank: 9/51 |
0.30180 (±0.00000) Rank: 34/51 |
0.40721 (±0.00000) Rank: 35/51 |
Tree | 8000.0 | — | 528.7 | 0.146 Rank: 1/51 |
0.018 Rank: 38/51 |
0.36203 (±0.00000) Rank: 34/51 |
0.47241 (±0.00000) Rank: 34/51 |
Avg | 8000.0 | — | 660.8 | 0.150 Rank: 1/51 |
0.028 Rank: 15/51 |
0.31624 (±0.00000) Rank: 34/51 |
0.41780 (±0.00000) Rank: 35/51 |
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) / ±0.13430 over 3 scenes)
Rank (per category): 34 (of 51)
Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
Lizard | 8000.0 | 851.16 | 84.17 Rank: 20/51 |
2906.86 Rank: 13/51 |
3.560 Rank: 5/51 |
14.17624 Rank: 27/51 |
0.50102 (±0.00000) Rank: 26/51 |
0.60128 (±0.00000) Rank: 25/51 |
Pond | 8000.0 | 1135.11 | 70.73 Rank: 12/51 |
2993.01 Rank: 11/51 |
3.162 Rank: 6/51 |
0.44359 Rank: 37/51 |
0.26867 (±0.00000) Rank: 17/51 |
0.29474 (±0.00000) Rank: 22/51 |
Tree | 8000.0 | 746.85 | 70.27 Rank: 31/51 |
1670.69 Rank: 13/51 |
3.108 Rank: 5/51 |
5.66765 Rank: 9/51 |
0.31358 (±0.00000) Rank: 34/51 |
0.34460 (±0.00000) Rank: 34/51 |
Avg | 8000.0 | 911.04 | 75.06 Rank: 22/51 |
2523.52 Rank: 14/51 |
3.277 Rank: 6/51 |
6.76249 Rank: 12/51 |
0.36109 (±0.00000) Rank: 34/51 |
0.41354 (±0.00000) Rank: 34/51 |
Google Urban dataset / Stereo track
mAA at 10 degrees: 0.16480 (±0.00000 over 1 run(s) / ±0.07428 over 17 scenes)
Rank (per category): 28 (of 51)
Scene | Features | Matches (raw) |
Matches (final) |
mAA(5o) | mAA(10o) |
Amsterdam (AMS) | 4334.3 | — | 409.2 | 0.07679 (±0.00000) Rank: 28/51 |
0.18304 (±0.00000) Rank: 28/51 |
Bangkok (BGK) | 3599.8 | — | 542.7 | 0.04069 (±0.00000) Rank: 25/51 |
0.10670 (±0.00000) Rank: 27/51 |
Barcelona (BCN) | 4610.3 | — | 273.8 | 0.02637 (±0.00000) Rank: 28/51 |
0.07326 (±0.00000) Rank: 28/51 |
Buenos Aires (BAR) | 4336.3 | — | 378.9 | 0.08037 (±0.00000) Rank: 28/51 |
0.15251 (±0.00000) Rank: 28/51 |
Cambridge (CAM) | 3294.7 | — | 314.6 | 0.09452 (±0.00000) Rank: 28/51 |
0.20103 (±0.00000) Rank: 28/51 |
Cannes (CAN) | 3459.3 | — | 422.7 | 0.18065 (±0.00000) Rank: 28/51 |
0.27742 (±0.00000) Rank: 28/51 |
Chicago (CHI) | 4249.8 | — | 125.9 | 0.05191 (±0.00000) Rank: 24/51 |
0.10362 (±0.00000) Rank: 22/51 |
Helsinki (HEL) | 3841.1 | — | 675.7 | 0.18144 (±0.00000) Rank: 28/51 |
0.32345 (±0.00000) Rank: 28/51 |
Madrid (MAD) | 4049.7 | — | 226.0 | 0.02796 (±0.00000) Rank: 27/51 |
0.06140 (±0.00000) Rank: 28/51 |
Mountain View (MTV) | 4192.5 | — | 230.1 | 0.03984 (±0.00000) Rank: 28/51 |
0.08867 (±0.00000) Rank: 28/51 |
New Orleans (NOR) | 3727.1 | — | 222.8 | 0.09704 (±0.00000) Rank: 28/51 |
0.15237 (±0.00000) Rank: 28/51 |
San Francisco (SF) | 3565.9 | — | 307.4 | 0.06667 (±0.00000) Rank: 27/51 |
0.15906 (±0.00000) Rank: 26/51 |
Singapore (SG) | 3334.7 | — | 265.7 | 0.08671 (±0.00000) Rank: 28/51 |
0.15570 (±0.00000) Rank: 28/51 |
Sydney (SYD) | 3714.6 | — | 247.8 | 0.06900 (±0.00000) Rank: 28/51 |
0.13406 (±0.00000) Rank: 28/51 |
Tokyo (TOK) | 3583.7 | — | 325.1 | 0.19510 (±0.00000) Rank: 28/51 |
0.30588 (±0.00000) Rank: 28/51 |
Toronto (TOR) | 3940.2 | — | 150.2 | 0.07040 (±0.00000) Rank: 28/51 |
0.13920 (±0.00000) Rank: 28/51 |
Zurich (ZRH) | 4461.8 | — | 340.6 | 0.11478 (±0.00000) Rank: 28/51 |
0.18419 (±0.00000) Rank: 28/51 |
Average (Avg) | 3899.7 | — | 321.1 | 0.08825 (±0.00000) Rank: 28/51 |
0.16480 (±0.00000) Rank: 28/51 |
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) / ±0.09788 over 17 scenes)
Rank (per category): 23 (of 51)
Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
AMS | 4334.3 | 229.01 | 81.75 Rank: 19/51 |
2027.50 Rank: 14/51 |
3.644 Rank: 11/51 |
21.15459 Rank: 9/51 |
0.22621 (±0.00000) Rank: 22/51 |
0.31004 (±0.00000) Rank: 23/51 |
BGK | 3599.8 | 339.72 | 93.33 Rank: 17/51 |
3215.99 Rank: 13/51 |
3.929 Rank: 1/51 |
10.53790 Rank: 3/51 |
0.01484 (±0.00000) Rank: 25/51 |
0.06008 (±0.00000) Rank: 24/51 |
BCN | 4610.3 | 148.37 | 77.83 Rank: 17/51 |
1078.65 Rank: 17/51 |
3.684 Rank: 3/51 |
24.24976 Rank: 8/51 |
0.06233 (±0.00000) Rank: 26/51 |
0.11211 (±0.00000) Rank: 26/51 |
BAR | 4336.3 | 193.53 | 68.64 Rank: 19/51 |
1363.32 Rank: 13/51 |
3.240 Rank: 8/51 |
12.33047 Rank: 12/51 |
0.04788 (±0.00000) Rank: 21/51 |
0.08857 (±0.00000) Rank: 21/51 |
CAM | 3294.7 | 175.90 | 75.79 Rank: 17/51 |
1532.17 Rank: 14/51 |
3.513 Rank: 5/51 |
21.89715 Rank: 15/51 |
0.12060 (±0.00000) Rank: 14/51 |
0.17531 (±0.00000) Rank: 14/51 |
CAN | 3459.3 | 220.03 | 69.13 Rank: 12/51 |
986.31 Rank: 13/51 |
3.514 Rank: 3/51 |
16.08550 Rank: 16/51 |
0.02841 (±0.00000) Rank: 18/51 |
0.06533 (±0.00000) Rank: 16/51 |
CHI | 4249.8 | 105.60 | 76.99 Rank: 20/51 |
431.04 Rank: 20/51 |
4.954 Rank: 7/51 |
44.87070 Rank: 15/51 |
0.06693 (±0.00000) Rank: 19/51 |
0.12988 (±0.00000) Rank: 17/51 |
HEL | 3841.1 | 470.51 | 89.99 Rank: 21/51 |
3512.60 Rank: 11/51 |
3.930 Rank: 4/51 |
11.42040 Rank: 4/51 |
0.21224 (±0.00000) Rank: 25/51 |
0.32618 (±0.00000) Rank: 25/51 |
MAD | 4049.7 | 131.68 | 78.57 Rank: 23/51 |
1319.44 Rank: 18/51 |
3.514 Rank: 1/51 |
3.45011 Rank: 3/51 |
0.03064 (±0.00000) Rank: 23/51 |
0.07561 (±0.00000) Rank: 23/51 |
MTV | 4192.5 | 154.55 | 81.84 Rank: 21/51 |
1399.35 Rank: 16/51 |
3.937 Rank: 1/51 |
13.65771 Rank: 9/51 |
0.11596 (±0.00000) Rank: 21/51 |
0.20731 (±0.00000) Rank: 21/51 |
NOR | 3727.1 | 142.19 | 62.74 Rank: 18/51 |
901.64 Rank: 13/51 |
3.209 Rank: 6/51 |
7.71879 Rank: 10/51 |
0.05172 (±0.00000) Rank: 23/51 |
0.08638 (±0.00000) Rank: 22/51 |
SF | 3565.9 | 176.15 | 82.09 Rank: 14/51 |
1398.73 Rank: 16/51 |
3.343 Rank: 10/51 |
15.55128 Rank: 16/51 |
0.08528 (±0.00000) Rank: 23/51 |
0.16564 (±0.00000) Rank: 23/51 |
SG | 3334.7 | 155.91 | 59.33 Rank: 23/51 |
714.67 Rank: 20/51 |
3.148 Rank: 10/51 |
12.72317 Rank: 8/51 |
0.02933 (±0.00000) Rank: 27/51 |
0.05878 (±0.00000) Rank: 27/51 |
SYD | 3714.6 | 131.90 | 72.53 Rank: 17/51 |
902.08 Rank: 15/51 |
3.474 Rank: 4/51 |
11.39633 Rank: 9/51 |
0.14430 (±0.00000) Rank: 16/51 |
0.21160 (±0.00000) Rank: 16/51 |
TOK | 3583.7 | 195.86 | 77.31 Rank: 23/51 |
1872.84 Rank: 15/51 |
3.616 Rank: 5/51 |
9.86043 Rank: 9/51 |
0.30779 (±0.00000) Rank: 23/51 |
0.37274 (±0.00000) Rank: 23/51 |
TOR | 3940.2 | 77.16 | 72.73 Rank: 19/51 |
699.66 Rank: 17/51 |
3.281 Rank: 6/51 |
13.31624 Rank: 15/51 |
0.04985 (±0.00000) Rank: 22/51 |
0.09072 (±0.00000) Rank: 23/51 |
ZRH | 4461.8 | 188.55 | 73.03 Rank: 20/51 |
1877.35 Rank: 13/51 |
3.134 Rank: 10/51 |
12.92367 Rank: 6/51 |
0.03754 (±0.00000) Rank: 21/51 |
0.06922 (±0.00000) Rank: 22/51 |
Avg | 3899.7 | 190.39 | 76.10 Rank: 20/51 |
1484.31 Rank: 15/51 |
3.592 Rank: 5/51 |
15.47907 Rank: 10/51 |
0.09599 (±0.00000) Rank: 23/51 |
0.15326 (±0.00000) Rank: 23/51 |
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 —