Update Readme and dataset
Browse files- README.md +31 -0
- eval_anywhere3D_aligned_total.json +0 -0
README.md
CHANGED
|
@@ -1,3 +1,34 @@
|
|
| 1 |
---
|
| 2 |
license: cc-by-4.0
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: cc-by-4.0
|
| 3 |
---
|
| 4 |
+
|
| 5 |
+
# Introduction to Anywhere3D-bench
|
| 6 |
+
|
| 7 |
+
eval_anywhere3D_aligned_total.json contains the all 2632 referring expressions-3D bounding box pairs from Anywhere3D benchmark.
|
| 8 |
+
|
| 9 |
+
The meaning of each key in the dictionary is as follows.The red color denotes the key-value pairs that directly related to expressions-3D bounding box pairs while the rest are helpful for annotation interface.
|
| 10 |
+
|
| 11 |
+
<span style="color:red">"grounding_level": grounding level of the expression, selected from one of the four values: ["area_level", "space_level", "object_level", "part_level"]. </span>
|
| 12 |
+
<span style="color:red">"expression_type": specifies the type of expression at each grounding level. </span>
|
| 13 |
+
"db_id": specifies the database ID in MongoDB during Human Annotation and Verification
|
| 14 |
+
"expressions_cnt_in_annotation_interface": specifies the expression number during Human Annotation and Verification
|
| 15 |
+
<span style="color:red">"datasetname": selected from one of the four values: ["scannet", "multiscan", "3RScan", "arkitscene_valid"].</span>
|
| 16 |
+
<span style="color:red">"scene_id": From simplicity during Human Annotation and Verification, we rename the scene_id in the original dataset</span>
|
| 17 |
+
<span style="color:red">"scene_name": The original scene_id of the four datasets</span>
|
| 18 |
+
<span style="color:red">"referring_expressions": referring expressions in Anywhere3D</span>
|
| 19 |
+
<span style="color:red">"box_width": The length of the ground-truth 3D bounding box along X-axis. </span>
|
| 20 |
+
<span style="color:red">"box_length" The length of the ground-truth 3D bounding box along Y-axis. </span>
|
| 21 |
+
<span style="color:red">"box_height" The length of the ground-truth 3D bounding box along Z-axis. </span>
|
| 22 |
+
<span style="color:red">"box_x": The center coordinate of the ground-truth 3D bounding box along X-axis. </span>
|
| 23 |
+
<span style="color:red">"box_y": The center coordinate of the ground-truth 3D bounding box along Y-axis. </span>
|
| 24 |
+
<span style="color:red">"box_z": The center coordinate of the ground-truth 3D bounding box along Z-axis. </span>
|
| 25 |
+
<span style="color:red">"box_rot_angle": The rotation angle of ground-truth 3D bounding box. </span>
|
| 26 |
+
|
| 27 |
+
"cylinder_diameter": The diameter of the distance measurement tool
|
| 28 |
+
"cylinder_length": The length of the distance measurement tool
|
| 29 |
+
"cylinder_x": The center coordinate of the distance measurement tool along X-axis
|
| 30 |
+
"cylinder_y": The center coordinate of the distance measurement tool along Y-axis
|
| 31 |
+
"cylinder_z": The center coordinate of the distance measurement tool along Z-axis
|
| 32 |
+
"cylinder_rot_angle": The rotation angle of distance measurement tool
|
| 33 |
+
"window_camera_position_x", "window_camera_position_y", "window_camera_position_z", "window_camera_quaternion_x", "window_camera_quaternion_y", "window_camera_quaternion_z", "window_camera_quaternion_w": The camera position and quaternion in annotation interface.
|
| 34 |
+
"window_camera_target_x", "window_camera_target_y", "window_camera_target_z": The camera target in annotation interface.
|
eval_anywhere3D_aligned_total.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|