Submission ID: 00635
MKDNet-indep
Processed: 20-12-13. Download link: sid-00635-sift8k_mkdnet_indep.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: (contact)
- Keypoint: sift
- Descriptor: mkdnet-indep (128 float32: 512 bytes)
- Number of features: 8000
- Summary: based on the paper [Explicit spatial encoding for deep local descriptors], trained on Liberty set from PhotoTourism dataset
- Paper: https://openaccess.thecvf.com/content_CVPR_2019/papers/Mukundan_Explicit_Spatial_Encoding_for_Deep_Local_Descriptors_CVPR_2019_paper.pdf
- Website: https://openaccess.thecvf.com/content_CVPR_2019/papers/Mukundan_Explicit_Spatial_Encoding_for_Deep_Local_Descriptors_CVPR_2019_paper.pdf
- Origin: Submission
- Flags: is_submission
Phototourism / Stereo track
mAA at 10 degrees: 0.51519 (±0.00069 over 3 run(s) / ±0.14801 over 9 scenes)
Rank (per category): 67 (of 147)
| Scene | Features | Matches (matcher) |
Matches (filter) |
Matches (final) |
Rep. @ 3 px. | MS @ 3 px. | mAA(5o) | mAA(10o) |
| bm | 7969.4 | 559.9 | 559.9 | 368.7 | 0.527 Rank: 30/147 |
0.939 Rank: 71/147 |
0.21213 (±0.00049) Rank: 60/147 |
0.32919 (±0.00017) Rank: 64/147 |
| fcs | 7874.2 | 672.2 | 672.2 | 452.4 | 0.443 Rank: 83/147 |
0.941 Rank: 28/147 |
0.62026 (±0.00146) Rank: 48/147 |
0.74109 (±0.00141) Rank: 46/147 |
| lms | 7770.9 | 465.6 | 465.6 | 240.6 | 0.468 Rank: 53/147 |
0.691 Rank: 76/147 |
0.52888 (±0.00109) Rank: 65/147 |
0.64734 (±0.00037) Rank: 63/147 |
| lb | 7953.6 | 431.2 | 431.2 | 249.3 | 0.474 Rank: 86/147 |
0.765 Rank: 54/147 |
0.42335 (±0.00049) Rank: 60/147 |
0.51353 (±0.00189) Rank: 76/147 |
| mc | 7824.0 | 576.5 | 576.5 | 378.7 | 0.524 Rank: 78/147 |
0.952 Rank: 71/147 |
0.37799 (±0.00049) Rank: 60/147 |
0.52540 (±0.00269) Rank: 78/147 |
| mr | 7797.2 | 518.4 | 518.4 | 371.2 | 0.457 Rank: 120/147 |
0.968 Rank: 62/147 |
0.29865 (±0.00143) Rank: 88/147 |
0.40556 (±0.00314) Rank: 87/147 |
| psm | 7809.6 | 304.5 | 304.5 | 175.5 | 0.385 Rank: 83/147 |
0.602 Rank: 78/147 |
0.14783 (±0.00181) Rank: 70/147 |
0.26111 (±0.00134) Rank: 69/147 |
| sf | 7896.6 | 715.1 | 715.1 | 471.2 | 0.480 Rank: 81/147 |
0.895 Rank: 19/147 |
0.48467 (±0.00312) Rank: 76/147 |
0.62451 (±0.00240) Rank: 68/147 |
| spc | 7869.2 | 554.9 | 554.9 | 321.0 | 0.487 Rank: 94/147 |
0.876 Rank: 46/147 |
0.43622 (±0.00099) Rank: 59/147 |
0.58903 (±0.00077) Rank: 57/147 |
| avg | 7862.7 | 533.1 | 533.1 | 336.5 | 0.472 Rank: 91/147 |
0.848 Rank: 50/147 |
0.39222 (±0.00095) Rank: 67/147 |
0.51519 (±0.00069) Rank: 67/147 |
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: 0.70004 (±0.00374 over 3 run(s) / ±0.13388 over 9 scenes)
Rank (per category): 62 (of 147)
| Scene | Features | Matches (input) |
RegistrationRatio (%) | Number of Landmarks |
Track Length | ATE | mAA(50) | mAA(100) |
| bm | 7969.4 | 552.69 | 99.90 Rank: 25/147 |
3565.35 Rank: 31/147 |
4.823 Rank: 109/147 |
0.51629 Rank: 79/147 |
0.44277 (±0.00707) Rank: 70/147 |
0.59360 (±0.00720) Rank: 71/147 |
| fcs | 7874.2 | 705.30 | 98.48 Rank: 54/147 |
5884.92 Rank: 29/147 |
4.392 Rank: 86/147 |
0.27171 Rank: 113/147 |
0.72679 (±0.00759) Rank: 63/147 |
0.78792 (±0.00743) Rank: 69/147 |
| lms | 7770.9 | 508.13 | 99.13 Rank: 65/147 |
4374.89 Rank: 26/147 |
4.325 Rank: 113/147 |
0.28981 Rank: 69/147 |
0.83006 (±0.00702) Rank: 50/147 |
0.88507 (±0.00726) Rank: 52/147 |
| lb | 7953.6 | 526.91 | 98.74 Rank: 3/147 |
3811.18 Rank: 26/147 |
4.508 Rank: 86/147 |
0.49080 Rank: 58/147 |
0.64367 (±0.00240) Rank: 70/147 |
0.73981 (±0.00602) Rank: 71/147 |
| mc | 7824.0 | 565.48 | 99.66 Rank: 74/147 |
4704.45 Rank: 47/147 |
4.494 Rank: 93/147 |
0.41237 Rank: 84/147 |
0.49807 (±0.00661) Rank: 81/147 |
0.64192 (±0.00714) Rank: 86/147 |
| mr | 7797.2 | 524.34 | 93.03 Rank: 106/147 |
4845.32 Rank: 70/147 |
4.098 Rank: 81/147 |
0.50651 Rank: 84/147 |
0.43377 (±0.01479) Rank: 70/147 |
0.54421 (±0.01726) Rank: 76/147 |
| psm | 7809.6 | 296.72 | 93.15 Rank: 59/147 |
4142.48 Rank: 39/147 |
3.089 Rank: 99/147 |
0.58473 Rank: 28/147 |
0.38264 (±0.00712) Rank: 72/147 |
0.47358 (±0.00998) Rank: 62/147 |
| sf | 7896.6 | 707.73 | 98.97 Rank: 25/147 |
6482.23 Rank: 53/147 |
4.383 Rank: 75/147 |
0.32328 Rank: 61/147 |
0.73655 (±0.00350) Rank: 35/147 |
0.82407 (±0.00333) Rank: 28/147 |
| spc | 7869.2 | 570.47 | 99.64 Rank: 40/147 |
5139.18 Rank: 26/147 |
4.269 Rank: 79/147 |
0.44477 Rank: 25/147 |
0.70982 (±0.00713) Rank: 40/147 |
0.81016 (±0.00871) Rank: 35/147 |
| avg | 7862.7 | 550.86 | 97.86 Rank: 52/147 |
4772.22 Rank: 31/147 |
4.265 Rank: 93/147 |
0.42670 Rank: 62/147 |
0.60046 (±0.00289) Rank: 61/147 |
0.70004 (±0.00374) Rank: 62/147 |
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 —
