File size: 6,368 Bytes
e6361b4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a7cebe5
e6361b4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
---
pretty_name: "MANUS-DexYCB: DexYCB-derived Multimodal Annotated Naturalistic Hand Understanding Dataset"
license: cc-by-nc-4.0
tags:
- image
- depth-estimation
- hand-pose
- hand-gesture
- 3d
- mano
- multimodal
- multi-view
- webdataset
- computer-vision
task_categories:
- image-to-image
- object-detection
- image-classification
configs:
- config_name: dexycb
  data_files:
  - split: test
    path: data/dexycb/test/*.tar
gated: false
---

# MANUS-DexYCB: DexYCB-derived Multimodal Annotated Naturalistic Hand Understanding Dataset

**MANUS-DexYCB** is the DexYCB-derived subset of the Multimodal Annotated Naturalistic Hand Understanding (MANUS) dataset family. It is released as a source-specific repository because MANUS subsets are governed by different upstream licenses.

This repository contains only the DexYCB-derived MANUS test split. HaGRID/HaGRIDv2-derived data is released separately as MANUS-HaGRID.

## Dataset Features in This Release

MANUS-DexYCB provides the following modalities where available:

- full-scene RGB image;
- full-scene depth map;
- hand-region RGB crop;
- hand-region depth map;
- 2D bounding boxes for hand regions;
- estimated MANO-style hand mesh annotations or related mesh metadata;
- rendered mesh views from six canonical viewpoints: front, rear, left, right, top, and bottom;
- primary and complementary mesh projection views where available;
- companion text prompts describing image content, provided as the split-level prompt file listed below;
- per-sample metadata recording source dataset, source id, split, available modalities, and license metadata.

Text prompts are provided in `DexYCB_test_prompt.txt` as a companion split-level file.

## Configs and Splits

| Config/Split | Samples | WebDataset shards | Size |
|---|---:|---:|---:|
| `dexycb/test` | 7,836 | 7 | 5.58 GB |

The canonical sample count is based on `DexYCB_test_OG_image_paths.txt`; other modalities are aligned to this OG image list.

## Data Format

Data is stored as uncompressed WebDataset tar shards. Files belonging to the same sample share the same prefix:

```text
dexycb_test_000001_processed_000001.json
dexycb_test_000001_processed_000001.full_rgb.png
dexycb_test_000001_processed_000001.full_depth.png
dexycb_test_000001_processed_000001.hand_rgb.png
dexycb_test_000001_processed_000001.hand_depth.png
dexycb_test_000001_processed_000001.hand_bbox.json
dexycb_test_000001_processed_000001.vertices.json
dexycb_test_000001_processed_000001.mesh_front.png
```

The per-sample `.json` member records source and license metadata, source id, split, available modalities, and the number of files available for that sample. See `docs/data_structure.md` and `docs/annotation_format.md` for details.

## Text Prompt Files

The following companion prompt file has been added at the repository root:

| File | Split | Records/Lines | Format | Notes |
|---|---|---:|---|---|
| `DexYCB_test_prompt.txt` | `dexycb/test` | 7,836 | TXT | One text prompt per line, aligned to the canonical DexYCB test OG image order. |

Prompt files are optional companion annotations and are not packed into the WebDataset shards. Use this file when text conditioning or image-content descriptions are needed.

## Metadata Files

- `metadata/dexycb_manifest.parquet`
- `metadata/dexycb_manifest.csv`
- `metadata/dexycb_manifest.jsonl`
- `metadata/source_licenses.csv`
- `metadata/release_schema.json`
- `metadata/release_report.json`
- `metadata/checksums.sha256`
- `DexYCB_test_prompt.txt`

The public manifest files do not contain local absolute source paths. They contain portable source and audit metadata, including sample id, source id, source license, available modalities, and shard path.

## Loading

See `docs/loading_examples.md`.

## Source Dataset Reference

This release is derived from DexYCB. Please acknowledge the upstream DexYCB dataset when using MANUS-DexYCB.

```bibtex
@inproceedings{chao2021dexycb,
  title={DexYCB: A benchmark for capturing hand grasping of objects},
  author={Chao, Yu-Wei and Yang, Wei and Xiang, Yu and Molchanov, Pavlo and Handa, Ankur and Tremblay, Jonathan and Narang, Yashraj S and Van Wyk, Karl and Iqbal, Umar and Birchfield, Stan and others},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={9044--9053},
  year={2021}
}
```

## Citation

If you use MANUS-DexYCB, please cite both the MANUS / HanDrawer work and the upstream DexYCB dataset.

### MANUS / HanDrawer

```bibtex
@misc{fu2025handrawer,
  title         = {HanDrawer: Leveraging Spatial Information to Render Realistic Hands Using a Conditional Diffusion Model in Single Stage},
  author        = {Qifan Fu and Xu Chen and Muhammad Asad and Shanxin Yuan and Changjae Oh and Gregory Slabaugh},
  year          = {2025},
  eprint        = {2503.02127},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  doi           = {10.48550/arXiv.2503.02127},
  url           = {https://arxiv.org/abs/2503.02127}
}
```

### DexYCB

Use the DexYCB citation listed in the source dataset reference above.

## License

MANUS-DexYCB is derived from DexYCB and is distributed under **Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)**. Commercial use is not permitted unless separately authorized by the DexYCB rights holders.

This repository does not grant any rights beyond those granted by the upstream DexYCB license. Users are responsible for complying with the upstream license, attribution requirements, change notices, and applicable laws.

## Intended Uses

MANUS-DexYCB is intended for research on hand-object interaction, hand-aware image generation, 2D/3D hand understanding, depth estimation, and multimodal hand representation learning.

## Out-of-Scope Uses

This dataset should not be used for biometric identification, person recognition, surveillance, or privacy-invasive applications.

## Privacy and Takedown

MANUS-DexYCB is derived from images of people interacting with objects. Some images may include identifiable visual context. If you are a rights holder or data subject and request removal of a sample, please contact the dataset maintainers through the repository contact channel.

## Related MANUS Repositories

- MANUS landing page: `MANUS`
- MANUS-HaGRID: HaGRID/HaGRIDv2-derived MANUS release

This repository contains only DexYCB-derived data.