File size: 4,973 Bytes
859ffcd 171af0b 859ffcd 171af0b 0fe8836 171af0b | 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 | ---
pretty_name: "EgoSuite-Open10K · EgoPro"
language:
- en
- zh
tags:
- robotics
- embodied-ai
- egocentric-video
- human-demonstration
- wrist-camera
- hand-pose
- body-pose
- semantic-annotation
- multimodal
viewer: false
configs:
- config_name: EgoPro
data_files:
- split: train
path: "manifests/EgoPro/*.parquet"
- config_name: EgoPro-motion
data_files:
- split: train
path: "manifests/EgoPro-motion/*.parquet"
- config_name: EgoProMax
default: true
data_files:
- split: train
path: "manifests/EgoProMax/*.parquet"
---
<div align="center">
<img src="./assets/hero.png" width="100%" alt="EgoPro synchronized head and wrist series" />
# EgoPro
### 1,000 hours of synchronized head-and-wrist egocentric activity for fine-grained manipulation.
[Collection](https://huggingface.co/collections/xuejf/egosuite-open10k-v2-test) · [EgoStandard](https://huggingface.co/datasets/xuejf/EgoStandard-v2-Test) · [EgoDemo](https://huggingface.co/datasets/xuejf/EgoDemo-v2-Test) · [Interactive Visualizer](https://huggingface.co/spaces/xuejf/EgoSuite-Visualizer-v2-Test) · [中文](./README_zh.md)
</div>
> **Private v2 test repository.** This card, schema, and visual design are ready for review. Audited Parquet manifests, synchronized media, approved license, checksums, and final citation are not present, so the Dataset Viewer stays disabled and the loading examples below are release-target examples.
## Hero SKU: EgoProMax
EgoProMax is the highest-supervision config in the Pro dual-view series. It binds synchronized head and wrist media, calibration metadata, 3D hand pose, 3D body pose, and V7 semantic references to one stable `clip_id`.
<img src="./assets/egopromax-multimodal-preview.png" width="100%" alt="EgoProMax synchronized wrist, pose, and semantic preview" />
```text
clip_id
|-- head_video_ref
|-- wrist_video_ref
|-- synchronization_ref
|-- hand_pose_3d_ref
|-- body_pose_3d_ref # EgoPro-motion and EgoProMax
`-- semantics_v7_ref # EgoProMax only
```
## Configs and hours
| Config | Hours | Capture and annotation contract | Role |
|---|---:|---|---|
| `EgoPro` | 750 | Synchronized head + wrist video + 3D hand pose | Base |
| `EgoPro-motion` | 200 | Base + 3D body pose | Motion |
| **`EgoProMax`** | **50** | Motion + V7 semantic references | **Hero SKU** |
| | **1,000** | | Repository total |
## EgoFull versus EgoProMax
| Capability | EgoFull | EgoProMax |
|---|---|---|
| Series | EgoStandard | EgoPro |
| Head video | Yes | Yes |
| Synchronized wrist video | No | Yes |
| 3D hand pose | Yes | Yes |
| 3D body pose | Yes | Yes |
| V7 semantic reference | Yes | Yes |
| Formal hours | 100 | 50 |
The two hero SKUs expose comparable high-level supervision depth, but they belong to different capture families. This test card does not claim that they contain the same source clips.
## Interactive dashboard
[Open EgoSuite-Visualizer](https://huggingface.co/spaces/xuejf/EgoSuite-Visualizer-v2-Test) and switch between `EgoFull Showcase` and `EgoProMax Showcase`. Hovering an environment updates the linked media panel and distribution detail.
<img src="./assets/dashboard-preview.png" width="100%" alt="EgoSuite Visualizer dashboard preview" />
Dataset Cards cannot execute the dashboard's JavaScript. The Space is the interactive surface; this card provides the stable summary and direct link.
## Target loading interface
```python
from datasets import load_dataset
promax = load_dataset(
"xuejf/EgoPro-v2-Test",
"EgoProMax",
split="train",
streaming=True,
)
sample = next(iter(promax))
print(sample["clip_id"])
print(sample["head_video_ref"])
print(sample["wrist_video_ref"])
print(sample["synchronization_ref"])
```
## Common manifest fields
| Field | Type | Meaning |
|---|---|---|
| `clip_id` | string | Stable anonymous clip key |
| `config_name` | string | Exact Hugging Face config name |
| `scene_family` | string | One of the seven environment families |
| `task_name` | string | Normalized task name |
| `duration_s` | float32 | Audited usable duration |
| `head_video_ref` | string | Head-view media reference |
| `wrist_video_ref` | string | Synchronized wrist-view reference |
| `synchronization_ref` | string | Timing and calibration reference |
| `hand_pose_3d_ref` | string | Frame-aligned hand-pose reference |
| `body_pose_3d_ref` | string, nullable | Body-pose reference |
| `semantics_v7_ref` | string, nullable | V7 semantic reference |
| `release_revision` | string | Immutable source revision |
| `sha256` | string | Primary media integrity checksum |
## Access, license, and citation
The approved release target is public discoverability with manual gated access. Pro access should require a named reviewer and a documented review policy.
The requested `cc-by-nc-sa-4.0` identifier is not declared because no approved license was supplied. Access approval does not override participant consent, privacy review, or final license restrictions.
|