vLAR commited on
Commit
f5237c5
·
verified ·
1 Parent(s): 0d1be35

Upload README.md

Browse files
Files changed (1) hide show
  1. README.md +9 -0
README.md CHANGED
@@ -557,6 +557,15 @@ Each entry in `transform_data` describes the pose of the object or component at
557
 
558
  </div>
559
 
 
 
 
 
 
 
 
 
 
560
  </div>
561
 
562
  <h3 id="cameras" style="font-size: 1.35em; margin-top: 0.75em; margin-bottom: 0.22em; line-height: 1.2; margin-left: 0;"><span style="display: inline-block; width: 1.08em;">•</span>Cameras</h3>
 
557
 
558
  </div>
559
 
560
+ <h3 id="trajectories" style="font-size: 1.35em; margin-top: 0.75em; margin-bottom: 0.22em; line-height: 1.2; margin-left: 0;"><span style="display: inline-block; width: 1.08em;">•</span>Trajectories</h3>
561
+ <div style="margin-left: 16px;">
562
+
563
+ Each sequence includes a same-named JSON file that records per-frame poses for selected dynamic objects in the world coordinate system. The file contains `sequence_info`, which stores basic sequence metadata such as frame count and `fps`, and `actors`, which stores object categories, asset index paths, actor types, and transformation data.
564
+
565
+ For `solid` actors, the object is treated as a single rigid body, and its pose sequence is stored in the actor-level `transform_data` field. For `interactable` actors, poses are computed separately for each component; therefore, the actor-level `transform_data` is empty, while each component stores its own poses under `components[component_name].transform_data`.
566
+
567
+ Each pose entry contains the frame index, timestamp, and transformation, including `location`, quaternion `rotation`, and `scale`.
568
+
569
  </div>
570
 
571
  <h3 id="cameras" style="font-size: 1.35em; margin-top: 0.75em; margin-bottom: 0.22em; line-height: 1.2; margin-left: 0;"><span style="display: inline-block; width: 1.08em;">•</span>Cameras</h3>