I87X: MASt3R-SLAM (base.yaml) on ScanVideos/Video2_T3; MasterSlam1 = no intrinsics, MasterSlam2 = ARCore intrinsics (--calib)

.gitattributes

@@ -1324,3 +1324,5 @@ I64/RTABMAP2/261105-111908.obj filter=lfs diff=lfs merge=lfs -text
 I64/RTABMAP3/261105-114053-Ip14.obj filter=lfs diff=lfs merge=lfs -text
 P64/MasterSlam2_ScanVideo1/scan.ply filter=lfs diff=lfs merge=lfs -text
 P64/MasterSlam1_ScanVideo1/scan.ply filter=lfs diff=lfs merge=lfs -text
+I87X/MasterSlam1_ScanVideo2/scan.ply filter=lfs diff=lfs merge=lfs -text
+I87X/MasterSlam2_ScanVideo2/scan.ply filter=lfs diff=lfs merge=lfs -text

I87X/MasterSlam1_ScanVideo2/config.yaml

@@ -0,0 +1,58 @@
+use_calib: False
+single_thread: False
+dataset:
+  subsample: 1
+  img_downsample: 1
+  center_principle_point: True
+
+matching:
+  max_iter: 10
+  lambda_init: 1e-8
+  convergence_thresh: 1e-6
+  dist_thresh: 1e-1 # distance in 3D space
+  radius: 3
+  dilation_max: 5 # Right now starts from max and iteratively decreases until 1
+
+tracking:
+  min_match_frac: 0.05
+  max_iters: 50
+  C_conf: 0.0
+  Q_conf: 1.5
+  rel_error: 1e-3
+  delta_norm: 1e-3
+  huber: 1.345
+  match_frac_thresh: 0.333
+  sigma_ray: 0.003
+  sigma_dist: 1e+1
+  sigma_pixel: 1.0
+  sigma_depth: 1e+1 # NOTE: log-depth!
+  sigma_point: 0.05
+  pixel_border: -10 # Only in calib (negative means allow pixels outside image up to that distance)
+  depth_eps: 1e-6 # Only in calib case
+  filtering_mode: weighted_pointmap # recent, first, best_score, weighted_pointmap, weighted_spherical, indep_conf
+  filtering_score: median # median, mean (only used for filtering_mode=best_score)
+
+local_opt:
+  pin: 1
+  window_size: 1e+6   
+  C_conf: 0.0
+  Q_conf: 1.5
+  min_match_frac: 0.1
+  pixel_border: -10 # Only in calib (negative means allow pixels outside image up to that distance)
+  depth_eps: 1e-6 # Only in calib case
+  max_iters: 10
+  sigma_ray: 0.003
+  sigma_dist: 1e+1
+  sigma_pixel: 1.0
+  sigma_depth: 1e+1 # NOTE: log-depth!
+  sigma_point: 0.05
+  delta_norm: 1e-8
+  use_cuda: True
+
+retrieval:
+  k: 3
+  min_thresh: 5e-3
+
+reloc:
+  min_match_frac: 0.3
+  strict: True

I87X/MasterSlam1_ScanVideo2/run.log

@@ -0,0 +1,93 @@
+/home/lehoangan/LHAN/mast3r/data/I87X_T3/scan
+{'use_calib': False, 'single_thread': False, 'dataset': {'subsample': 1, 'img_downsample': 1, 'center_principle_point': True}, 'matching': {'max_iter': 10, 'lambda_init': 1e-08, 'convergence_thresh': 1e-06, 'dist_thresh': 0.1, 'radius': 3, 'dilation_max': 5}, 'tracking': {'min_match_frac': 0.05, 'max_iters': 50, 'C_conf': 0.0, 'Q_conf': 1.5, 'rel_error': 0.001, 'delta_norm': 0.001, 'huber': 1.345, 'match_frac_thresh': 0.333, 'sigma_ray': 0.003, 'sigma_dist': 10.0, 'sigma_pixel': 1.0, 'sigma_depth': 10.0, 'sigma_point': 0.05, 'pixel_border': -10, 'depth_eps': 1e-06, 'filtering_mode': 'weighted_pointmap', 'filtering_score': 'median'}, 'local_opt': {'pin': 1, 'window_size': 1000000.0, 'C_conf': 0.0, 'Q_conf': 1.5, 'min_match_frac': 0.1, 'pixel_border': -10, 'depth_eps': 1e-06, 'max_iters': 10, 'sigma_ray': 0.003, 'sigma_dist': 10.0, 'sigma_pixel': 1.0, 'sigma_depth': 10.0, 'sigma_point': 0.05, 'delta_norm': 1e-08, 'use_cuda': True}, 'retrieval': {'k': 3, 'min_thresh': 0.005}, 'reloc': {'min_match_frac': 0.3, 'strict': True}}
+... loading model from checkpoints/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric.pth
+/home/lehoangan/LHAN/mast3r/MASt3R-SLAM/thirdparty/mast3r/mast3r/model.py:24: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
+  ckpt = torch.load(model_path, map_location='cpu')
+instantiating : AsymmetricMASt3R(enc_depth=24, dec_depth=12, enc_embed_dim=1024, dec_embed_dim=768, enc_num_heads=16, dec_num_heads=12, pos_embed='RoPE100',img_size=(512, 512), head_type='catmlp+dpt', output_mode='pts3d+desc24', depth_mode=('exp', -inf, inf), conf_mode=('exp', 1, inf), patch_embed_cls='PatchEmbedDust3R', two_confs=True, desc_conf_mode=('exp', 0, inf), landscape_only=False)
+<All keys matched successfully>
+Loading retrieval model from checkpoints/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric_retrieval_trainingfree.pth
+/home/lehoangan/LHAN/mast3r/MASt3R-SLAM/thirdparty/mast3r/mast3r/retrieval/processor.py:70: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
+  ckpt = torch.load(modelname, 'cpu')  # TODO from pretrained to download it automatically
+/home/lehoangan/LHAN/mast3r/MASt3R-SLAM/mast3r_slam/frame.py:116: UserWarning: The given NumPy array is not writable, and PyTorch does not support non-writable tensors. This means writing to this tensor will result in undefined behavior. You may want to copy the array to protect its data or make it writable before converting it to a tensor. This type of warning will be suppressed for the rest of this program. (Triggered internally at /opt/conda/conda-bld/pytorch_1729647429097/work/torch/csrc/utils/tensor_numpy.cpp:206.)
+  uimg = torch.from_numpy(img["unnormalized_img"]) / 255.0
+FPS: 8.640121356685011
+FPS: 9.060991890668788
+FPS: 9.258809561739502
+FPS: 9.30855702097904
+FPS: 9.342082210334143
+Database retrieval 2 :  {0}
+FPS: 9.360102502877274
+FPS: 9.419028855520374
+FPS: 9.434152246730337
+FPS: 9.472562283492426
+FPS: 9.48588939395214
+Database retrieval 3 :  {0}
+FPS: 9.433322319432957
+FPS: 9.435805134001011
+FPS: 9.444843986430934
+FPS: 9.44991854147795
+FPS: 9.428561580434693
+FPS: 9.436365572218465
+FPS: 9.4356452960193
+FPS: 9.437316043780312
+FPS: 9.4356806850056
+FPS: 9.442766823422096
+FPS: 9.449006664631693
+FPS: 9.46327105519892
+FPS: 9.471254749106961
+FPS: 9.480566628490411
+Database retrieval 6 :  {3, 4}
+FPS: 9.46950189484681
+FPS: 9.483980953124401
+Database retrieval 7 :  {3, 4}
+FPS: 9.472545433103797
+FPS: 9.475964961933155
+FPS: 9.483812266728904
+FPS: 9.494100802014549
+FPS: 9.499581364332544
+Database retrieval 8 :  {0, 2}
+FPS: 9.490432682433783
+Database retrieval 9 :  {0, 1}
+FPS: 9.48628962733767
+FPS: 9.494942587512039
+FPS: 9.490830621806513
+Database retrieval 10 :  {5, 6}
+FPS: 9.48171971652284
+FPS: 9.487423200804697
+FPS: 9.484805454747278
+Database retrieval 11 :  {4, 5, 6}
+FPS: 9.466888675324334
+FPS: 9.472424390706582
+Database retrieval 12 :  {4, 5}
+FPS: 9.458997561539428
+FPS: 9.460091120751324
+FPS: 9.451890311536706
+FPS: 9.447949261796541
+Database retrieval 14 :  {11, 12, 4}
+FPS: 9.440829875416005
+Database retrieval 15 :  {3, 6, 7}
+FPS: 9.42981643317176
+FPS: 9.427228706080845
+FPS: 9.428687029304749
+Database retrieval 16 :  {4, 12, 14}
+FPS: 9.414644021280857
+Database retrieval 17 :  {12, 14}
+FPS: 9.403626631761467
+FPS: 9.406342180934317
+Database retrieval 18 :  {16, 14}
+FPS: 9.39892287380275
+Database retrieval 19 :  {3, 15, 7}
+FPS: 9.390766530721928
+FPS: 9.393771974463663
+FPS: 9.393850717461596
+FPS: 9.38652713397729
+FPS: 9.388118825172848
+Database retrieval 23 :  {16, 18, 7}
+FPS: 9.38215341589449
+Database retrieval 24 :  {19, 15, 7}
+FPS: 9.374459033826174
+Database retrieval 25 :  {8, 2}
+FPS: 9.372833138692384
+FPS: 9.378969741898134
+FPS: 9.385985733764004
+done

I87X/MasterSlam1_ScanVideo2/scan.ply

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8cf03018924fda086ee923c3cb5a02310a09ccd924132084061886312f2cf58
+size 48476386

I87X/MasterSlam1_ScanVideo2/scan.txt

@@ -0,0 +1,26 @@
+0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0
+4.233333110809326 -0.4181569814682007 0.009866035543382168 0.16344276070594788 -0.05217243731021881 -0.08688927441835403 -0.037785861641168594 0.9941333532333374
+5.4666666984558105 -0.5514125823974609 -0.039957594126462936 0.3437071144580841 -0.05206074193120003 0.1765880286693573 0.0939127579331398 0.9784095287322998
+11.0 -0.6542526483535767 -0.11795662343502045 0.4141215980052948 0.13113710284233093 0.16430985927581787 0.040454715490341187 0.9768149852752686
+12.033333778381348 -0.6581905484199524 -0.33267706632614136 0.3308194577693939 0.2819123864173889 0.12578226625919342 -0.010338393971323967 0.951103687286377
+14.566666603088379 -0.745076060295105 -0.21932290494441986 0.1839430034160614 0.27161043882369995 -0.1600014567375183 0.003920601215213537 0.949006199836731
+24.96666717529297 -0.7301332354545593 -0.16347457468509674 0.3524002432823181 0.1509547382593155 0.01069391518831253 0.012786071747541428 0.9883999228477478
+26.899999618530273 -0.6053348183631897 -0.13749772310256958 0.2542928457260132 0.15418703854084015 0.23783735930919647 0.04516297206282616 0.9579253792762756
+31.03333282470703 -0.6286792159080505 -0.05136174336075783 0.31142672896385193 -0.031555116176605225 0.21509188413619995 0.09479700773954391 0.9714698791503906
+32.93333435058594 -0.7656772136688232 -0.056109461933374405 0.2555813789367676 -0.00752122700214386 -0.04787706211209297 0.0007414862629957497 0.9988245368003845
+35.29999923706055 -0.7966336607933044 -0.14883311092853546 0.28054460883140564 0.15233740210533142 -0.12754717469215393 -0.02897888980805874 0.9796352386474609
+38.599998474121094 -0.7244316935539246 -0.19419404864311218 0.30718791484832764 0.25920847058296204 0.04496121034026146 -0.015122681856155396 0.9646559357643127
+40.63333511352539 -0.6149085760116577 -0.21851344406604767 0.21966871619224548 0.36301809549331665 0.1661406010389328 -0.009067433886229992 0.9168058633804321
+42.266666412353516 -0.6374486684799194 -0.2903498411178589 0.24803978204727173 0.5826855301856995 0.17106859385967255 -0.061942730098962784 0.7920714616775513
+44.33333206176758 -0.6253734230995178 -0.18611446022987366 0.21065232157707214 0.3626643717288971 0.18231233954429626 -0.0014114606892690063 0.9139125943183899
+45.46666717529297 -0.5730672478675842 -0.13249462842941284 0.3231814205646515 0.17140132188796997 0.2162761092185974 0.029078150168061256 0.9607300758361816
+48.599998474121094 -0.3242822587490082 -0.2555272877216339 0.23517754673957825 0.2728402018547058 0.25311729311943054 0.014786822721362114 0.9280474185943604
+49.56666564941406 -0.35771316289901733 -0.2577853202819824 0.2507112920284271 0.4368840456008911 0.24760262668132782 -0.03192336857318878 0.8641796708106995
+51.733333587646484 -0.310634583234787 -0.246626615524292 0.3019798696041107 0.2779374420642853 0.25476473569869995 0.023658331483602524 0.9258976578712463
+52.900001525878906 -0.3240581452846527 -0.15111394226551056 0.3254895508289337 0.08363674581050873 0.2907800078392029 0.08923257142305374 0.9489417672157288
+54.233333587646484 -0.28633591532707214 -0.09975221753120422 0.21008753776550293 0.042940132319927216 0.5476610064506531 0.16308347880840302 0.8195285797119141
+55.86666488647461 -0.238691508769989 -0.09145725518465042 0.11041518300771713 0.06722220033407211 0.7227810025215149 0.1963248997926712 0.6591855883598328
+56.66666793823242 -0.27497631311416626 -0.16130566596984863 0.15452419221401215 0.17643049359321594 0.6598373055458069 0.10188401490449905 0.723261296749115
+58.0 -0.3521493673324585 -0.2050265520811081 0.26020094752311707 0.19306868314743042 0.42528489232063293 0.08495861291885376 0.8801360130310059
+58.766666412353516 -0.4083215296268463 -0.1610308736562729 0.36027976870536804 0.1075100228190422 0.27970534563064575 0.06007911264896393 0.9521539211273193
+59.733333587646484 -0.4632929563522339 -0.09547151625156403 0.333832710981369 -0.054752416908741 0.20991140604019165 0.0568837933242321 0.9745272994041443

I87X/MasterSlam2_ScanVideo2/calib_arcore.yaml

@@ -0,0 +1,4 @@
+# From /home/lehoangan/LHAN/RoSE/I87X/ScanVideos/Video2_T3/intrinsics.json (ARCore, sensor-oriented 1920x1080 image)
+width: 1920
+height: 1080
+calibration: [1273.68310546875, 1288.348388671875, 965.7686767578125, 527.375244140625]

I87X/MasterSlam2_ScanVideo2/config.yaml

@@ -0,0 +1,58 @@
+use_calib: False
+single_thread: False
+dataset:
+  subsample: 1
+  img_downsample: 1
+  center_principle_point: True
+
+matching:
+  max_iter: 10
+  lambda_init: 1e-8
+  convergence_thresh: 1e-6
+  dist_thresh: 1e-1 # distance in 3D space
+  radius: 3
+  dilation_max: 5 # Right now starts from max and iteratively decreases until 1
+
+tracking:
+  min_match_frac: 0.05
+  max_iters: 50
+  C_conf: 0.0
+  Q_conf: 1.5
+  rel_error: 1e-3
+  delta_norm: 1e-3
+  huber: 1.345
+  match_frac_thresh: 0.333
+  sigma_ray: 0.003
+  sigma_dist: 1e+1
+  sigma_pixel: 1.0
+  sigma_depth: 1e+1 # NOTE: log-depth!
+  sigma_point: 0.05
+  pixel_border: -10 # Only in calib (negative means allow pixels outside image up to that distance)
+  depth_eps: 1e-6 # Only in calib case
+  filtering_mode: weighted_pointmap # recent, first, best_score, weighted_pointmap, weighted_spherical, indep_conf
+  filtering_score: median # median, mean (only used for filtering_mode=best_score)
+
+local_opt:
+  pin: 1
+  window_size: 1e+6   
+  C_conf: 0.0
+  Q_conf: 1.5
+  min_match_frac: 0.1
+  pixel_border: -10 # Only in calib (negative means allow pixels outside image up to that distance)
+  depth_eps: 1e-6 # Only in calib case
+  max_iters: 10
+  sigma_ray: 0.003
+  sigma_dist: 1e+1
+  sigma_pixel: 1.0
+  sigma_depth: 1e+1 # NOTE: log-depth!
+  sigma_point: 0.05
+  delta_norm: 1e-8
+  use_cuda: True
+
+retrieval:
+  k: 3
+  min_thresh: 5e-3
+
+reloc:
+  min_match_frac: 0.3
+  strict: True