ZGTZ-2026 commited on
Commit
e3cc10d
·
verified ·
1 Parent(s): 0fa6f93

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +49 -0
README.md CHANGED
@@ -1,3 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  <div style="background-color: #D72222; padding: 0 20px 30px 20px;">
2
 
3
  <div style="background:#D72222; padding:80px 20px; margin:-40px -20px; text-align:center; color:white;">
 
1
+ ---
2
+ # Dataset Card for Multi-View Continuous 4D Human Motion Dataset
3
+ language:
4
+ - en
5
+ - zh
6
+ license: unknown
7
+ size_categories:
8
+ - 10000h+
9
+ task_categories:
10
+ - computer-vision
11
+ - 3d-vision
12
+ - video-processing
13
+ task_ids:
14
+ - human-pose-estimation
15
+ - action-recognition
16
+ - 3d-point-cloud-processing
17
+ - multi-view-vision
18
+ dataset_info:
19
+ features:
20
+ - name: scene
21
+ dtype: string
22
+ - name: number_of_people
23
+ dtype: int64
24
+ - name: action
25
+ dtype: string
26
+ - name: smpl_parameters
27
+ dtype: dict
28
+ - name: 3d_point_cloud
29
+ dtype: array
30
+ - name: video_quality
31
+ dtype: string
32
+ - name: acquisition_method
33
+ dtype: string
34
+ creator: 中广天择
35
+ contact: unknown
36
+ tags:
37
+ - 3D human pose
38
+ - multi-person action
39
+ - multi-view video
40
+ - 4D human motion
41
+ - SMPL
42
+ - 3D point cloud
43
+ - 1080P HD video
44
+ description: |
45
+ Large-Scale 1080P Multi-View Human Motion Data for AI Research & Commercial Applications.
46
+ This dataset includes 10,000+ total hours of shooting, 2,000+ real-world scenes, 6,000+ subject instances,
47
+ covering 800+ hours of multi-person interactions, 2,200+ hours of dual-person actions, 7,000+ hours of single-person actions.
48
+ ---
49
+
50
  <div style="background-color: #D72222; padding: 0 20px 30px 20px;">
51
 
52
  <div style="background:#D72222; padding:80px 20px; margin:-40px -20px; text-align:center; color:white;">