This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
(disk-cpu) trulls ~/src/disk $ colmap exhaustive_matcher --database_path out/xxx-10fps-disk-16d/database2.db --SiftMatching.use_gpu 0 | |
============================================================================== | |
Exhaustive feature matching | |
============================================================================== | |
Matching block [1/31, 1/31] in 19.855s | |
Matching block [1/31, 2/31] in 0.016s | |
Matching block [1/31, 3/31] in 0.010s | |
Matching block [1/31, 4/31] in 0.017s |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
root@e6c5456fa8b4:/home/trulls/src/disco-code/points# python extract_benchmark.py | |
[['2020-05-25/768-depth', 504, True]] | |
Processing /home/trulls/imw-2020/sagrada_familia and saving to out/2020-05-25/768-depth-imsize-504-nms-5/sagrada_familia | |
0%| | 0/100 [00:00<?, ?it/s] | |
Traceback (most recent call last): | |
File "detect.py", line 144, in <module> | |
described_samples = extract(extraction_method, dataset, args.h5_path) | |
File "detect.py", line 101, in extract |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Processing /home/trulls/imw-2020/milan_cathedral and saving to out/2020-05-25/768-depth-imsize-960/milan_cathedral | |
0%| | 0/100 [00:00<?, ?it/s]ERROR: Unexpected bus error encountered in worker. This might be caused by insufficient shared memory (shm). | |
ERROR: Unexpected bus error encountered in worker. This might be caused by insufficient shared memory (shm). | |
Traceback (most recent call last): | |
File "/usr/lib/python3.8/multiprocessing/queues.py", line 239, in _feed | |
obj = _ForkingPickler.dumps(obj) | |
File "/usr/lib/python3.8/multiprocessing/reduction.py", line 51, in dumps | |
cls(buf, protocol).dump(obj) | |
File "/usr/local/lib/python3.8/dist-packages/torch/multiprocessing/reductions.py", line 333, in reduce_storage | |
fd, size = storage._share_fd_() |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
WAF-N16 (8k) | |
CV2-RANSAC, Best : 0.75 -> 0.4290487232505875 | |
NODEGENSAC, Best : 0.5 -> 0.45799136562326015 | |
DEGENSAC, Best : 1 -> 0.5441916837751245 | |
GCRANSAC, Best : 1.25 -> 0.5153767755827378 | |
MAGSAC, Best : 2 -> 0.5071992512758609 | |
WASF-N16 (8k) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
NODEGENSAC, CV-AKAZE -> Best mAP=0.3438 at th=0.25 | |
NODEGENSAC, CV-DoG/GeoDesc -> Best mAP=0.5078 at th=0.2 | |
NODEGENSAC, CV-DoG/HardNet -> Best mAP=0.6288 at th=0.25 | |
NODEGENSAC, CV-DoG/LogPolarDesc -> Best mAP=0.5732 at th=0.2 | |
NODEGENSAC, CV-DoG/SOSNet -> Best mAP=0.5701 at th=0.25 | |
NODEGENSAC, CV-ORB -> Best mAP=0.2359 at th=0.75 | |
NODEGENSAC, CV-RootSIFT -> Best mAP=0.5638 at th=0.25 | |
NODEGENSAC, CV-SIFT -> Best mAP=0.5386 at th=0.25 | |
NODEGENSAC, CV-SURF -> Best mAP=0.3052 at th=0.75 | |
NODEGENSAC, ContextDesc -> Best mAP=0.6058 at th=0.5 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
diff --git a/run.py b/run.py | |
index 2d1fffd5..8f7d66ae 100644 | |
--- a/run.py | |
+++ b/run.py | |
@@ -165,7 +165,10 @@ def main(cfg): | |
match_jobs = create_eval_jobs(feature_jobs, "match", cfg, job_dict) | |
# Filter | |
- match_inlier_jobs = create_eval_jobs(match_jobs, "filter", cfg, job_dict) | |
+ if cfg_m.refine_inliers: |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import numpy as np | |
from third_party.colmap.scripts.python.read_write_model import read_model, qvec2rotmat | |
src = '../rebuttal/results-colmap/' | |
considered = [5, 10, 25, 50, 100, 150, 200, 250, 300, 400, 500, 600] | |
cameras = {} | |
images = {} | |
names = {} |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
(...) | |
-- Packing "phototourism"/"united_states_capitol"/stereo, fold: 1/3, metric: avg_num_keypoints [0.69 s] | |
-- Packing "phototourism"/"united_states_capitol"/stereo, fold: 1/3, metric: num_inliers [28.87 s] | |
-- Packing "phototourism"/"united_states_capitol"/stereo, fold: 1/3, metric: matching_scores_depth_projection [32.78 s] | |
-- Packing "phototourism"/"united_states_capitol"/stereo, fold: 1/3, metric: repeatability [6.53 s] | |
-- Packing "phototourism"/"united_states_capitol"/stereo, fold: 1/3, metric: qt_map [6.49 s] | |
-- Packing "phototourism"/"united_states_capitol"/stereo, fold: 1/3, metric: timings [0.06 s] | |
-- Packing "phototourism"/"united_states_capitol"/stereo, fold: 2/3, metric: avg_num_keypoints [0.06 s] | |
-- Packing "phototourism"/"united_states_capitol"/stereo, fold: 2/3, metric: num_inliers [30.00 s] | |
-- Packing "phototourism"/"united_states_capitol"/stereo, fold: 2/3, metric: matching_scores_depth_projection [33.06 s] |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
-- computing filter | |
WARNING: ./jobs/5e2d15ede4bfc131456db7c0063449b99f62fbe4bbb159ab2a9b835454958907 already exists! | |
/home/trulls_google_com/miniconda3/envs/opencv-3.7/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:516: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'. | |
_np_qint8 = np.dtype([("qint8", np.int8, 1)]) | |
/home/trulls_google_com/miniconda3/envs/opencv-3.7/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:517: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'. | |
_np_quint8 = np.dtype([("quint8", np.uint8, 1)]) | |
/home/trulls_google_com/miniconda3/envs/opencv-3.7/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:518: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
STEREO ON TEST | |
mAPs -> [0.5286 0.5233 0.5125 0.508 0.4971 0.4921 0.4856 0.4787 0.47 0.4659 | |
0.4653 0.4643 0.4608 0.4556 0.4489 0.4204 0.3142 0.2669 0.2481 0.2253 | |
0.1836 0.1723 0.1639] | |
consecutive diffs -> [0.0053 0.0108 0.0045 0.0109 0.005 0.0065 0.0069 0.0087 0.0041 0.0006 | |
0.001 0.0035 0.0052 0.0067 0.0285 0.1062 0.0473 0.0188 0.0228 0.0417 | |
0.0113 0.0084] | |
Stereo on "test" (23 methods): Mean -> 0.4022, min -> 0.1639, max -> 0.5286, std -> 0.1230, smallest delta -> 0.0006, mean delta -> 0.0166 |
NewerOlder