RayZuo commited on
Commit
2215715
·
verified ·
1 Parent(s): a5a292f
Files changed (2) hide show
  1. Croissant.json +1 -1
  2. README.md +123 -0
Croissant.json CHANGED
@@ -55,7 +55,7 @@
55
  {
56
  "@type": "cr:FileObject",
57
  "@id": "data-zip",
58
- "contentUrl": "https://huggingface.co/datasets/RayZuo/CoRoleHOI/resolve/main/data.zip",
59
  "encodingFormat": "application/zip"
60
  }
61
  ],
 
55
  {
56
  "@type": "cr:FileObject",
57
  "@id": "data-zip",
58
+ "contentUrl": "https://huggingface.co/datasets/RayZuo/CoRoleHOI/resolve/main/CoRoleHOI.zip",
59
  "encodingFormat": "application/zip"
60
  }
61
  ],
README.md CHANGED
@@ -1,3 +1,126 @@
1
  ---
 
2
  license: cc-by-4.0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ pretty_name: CoRoleHOI
3
  license: cc-by-4.0
4
+ task_categories:
5
+ - other
6
+ task_ids:
7
+ - text-to-motion
8
+ - human-object-interaction
9
+ language:
10
+ - en
11
+ tags:
12
+ - human-object-interaction
13
+ - multi-person
14
+ - role-asymmetry
15
+ - smpl-x
16
+ - 3d
17
+ - embodied-ai
18
+ - motion-generation
19
+ annotations_creators:
20
+ - expert-generated
21
+ source_datasets:
22
+ - original
23
+ size_categories:
24
+ - 10K<n<100K
25
  ---
26
+
27
+ # CoRoleHOI Dataset
28
+
29
+ ## Overview
30
+
31
+ **CoRoleHOI** is a dataset for **role-asymmetric multi-person multi-object Human–Object Interaction (HOI)**.
32
+
33
+ It is designed to support research in:
34
+ - Text-driven motion generation
35
+ - Multi-person interaction modeling
36
+ - Human-object interaction understanding
37
+ - Embodied AI and robotics
38
+
39
+ The dataset integrates multi-modal signals:
40
+ - SMPL-X human motion sequences
41
+ - Object trajectories (4×4 transformation matrices)
42
+ - Fine-grained contact annotations
43
+ - Structured textual prompts
44
+ - Object meshes and SDF representations
45
+
46
+ ---
47
+
48
+ ## Key Contributions
49
+
50
+ - Role-asymmetric interaction modeling (e.g., giver → receiver)
51
+ - Multi-person coordination scenarios
52
+ - Multi-object interaction support
53
+ - Fine-grained contact supervision
54
+ - Aligned text–motion–object representation
55
+
56
+ ---
57
+
58
+ ## Dataset Structure
59
+
60
+ The dataset is distributed primarily as a compressed archive:
61
+
62
+ CoRoleHOI/
63
+ ├── data.zip
64
+ ├── README.md
65
+ ├── croissant.json
66
+
67
+ Contents of data.zip:
68
+
69
+ data/
70
+ ├── smplx_npz/
71
+ ├── obj_transmat_npy/
72
+ ├── contact_labels_w_semantics_npy_files/
73
+ ├── text_anno_json_data/
74
+ ├── human_metadata/
75
+ ├── obj_metadata/
76
+ ├── captured_objects/
77
+ ├── rest_object_sdf_256_npy_files/
78
+ ├── augmented_prompts.json
79
+ ├── prompt_dict.json
80
+ ├── split_prompts.json
81
+
82
+ ---
83
+
84
+ ## Data Modalities
85
+
86
+ Each interaction sequence includes:
87
+
88
+ | Modality | Description |
89
+ | -------------- | ---------------------------------------- |
90
+ | Human Motion | SMPL-X parameters |
91
+ | Object Motion | 4×4 transformation matrices |
92
+ | Contact Labels | Binary / semantic contact |
93
+ | Text | Natural language interaction description |
94
+ | Metadata | Human & object attributes |
95
+
96
+ ---
97
+
98
+ ## Text Annotations
99
+
100
+ The dataset provides structured language supervision:
101
+
102
+ - augmented_prompts.json
103
+ Multiple paraphrases per interaction
104
+
105
+ - prompt_dict.json
106
+ Prompt ID to canonical text mapping
107
+
108
+ - split_prompts.json
109
+ Train / validation / test splits
110
+
111
+ ---
112
+
113
+ ## Example Interaction
114
+
115
+ Person A passes the printer to Person B.
116
+
117
+ ---
118
+
119
+ ## Download
120
+
121
+ Dataset page:
122
+ https://huggingface.co/datasets/RayZuo/CoRoleHOI
123
+
124
+ Direct download:
125
+ https://huggingface.co/datasets/RayZuo/CoRoleHOI/resolve/main/CoRoleHOI.zip
126
+