Hugging Face's logo

Datasets:
RayZuo
/
CoRoleHOI

Tasks:
Other
Modalities:
3D
Languages:
English
Size:
10K<n<100K
Tags:
human-object-interaction
multi-person
role-asymmetry
smpl-x
3d
embodied-ai
License:
Dataset card Files
xet
 Community
CoRoleHOI
File size: 2,837 Bytes
154da10
2215715
154da10
2215715
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
154da10
2215715
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
---

pretty_name: CoRoleHOI
license: cc-by-4.0
task_categories:
- other
task_ids:
- text-to-motion
- human-object-interaction
language:
- en
tags:
- human-object-interaction
- multi-person
- role-asymmetry
- smpl-x
- 3d
- embodied-ai
- motion-generation
annotations_creators:
- expert-generated
source_datasets:
- original
size_categories:
- 10K<n<100K
---


# CoRoleHOI Dataset

## Overview

**CoRoleHOI** is a dataset for **role-asymmetric multi-person multi-object Human–Object Interaction (HOI)**.

It is designed to support research in:
- Text-driven motion generation
- Multi-person interaction modeling
- Human-object interaction understanding
- Embodied AI and robotics

The dataset integrates multi-modal signals:
- SMPL-X human motion sequences
- Object trajectories (4×4 transformation matrices)
- Fine-grained contact annotations
- Structured textual prompts
- Object meshes and SDF representations

---

## Key Contributions

- Role-asymmetric interaction modeling (e.g., giver → receiver)
- Multi-person coordination scenarios
- Multi-object interaction support
- Fine-grained contact supervision
- Aligned text–motion–object representation

---

## Dataset Structure

The dataset is distributed primarily as a compressed archive:

CoRoleHOI/
├── data.zip
├── README.md
├── croissant.json

Contents of data.zip:

data/
├── smplx_npz/

├── obj_transmat_npy/

├── contact_labels_w_semantics_npy_files/
├── text_anno_json_data/

├── human_metadata/
├── obj_metadata/

├── captured_objects/
├── rest_object_sdf_256_npy_files/

├── augmented_prompts.json
├── prompt_dict.json

├── split_prompts.json

---

## Data Modalities

Each interaction sequence includes:

| Modality       | Description                              |
| -------------- | ---------------------------------------- |
| Human Motion   | SMPL-X parameters                        |
| Object Motion  | 4×4 transformation matrices              |
| Contact Labels | Binary / semantic contact                |
| Text           | Natural language interaction description |
| Metadata       | Human & object attributes                |

---

## Text Annotations

The dataset provides structured language supervision:

- augmented_prompts.json  

  Multiple paraphrases per interaction



- prompt_dict.json  
  Prompt ID to canonical text mapping

- split_prompts.json  

  Train / validation / test splits



---



## Example Interaction



Person A passes the printer to Person B.



---



## Download



Dataset page:  

https://huggingface.co/datasets/RayZuo/CoRoleHOI



Direct download:  

https://huggingface.co/datasets/RayZuo/CoRoleHOI/resolve/main/CoRoleHOI.zip