Update README.md
Browse files
README.md
CHANGED
|
@@ -27,10 +27,10 @@ dataset_info:
|
|
| 27 |
---
|
| 28 |
|
| 29 |
# ChingMu 1000-Hour Embodied Motion Dataset
|
| 30 |
-
|
| 31 |
|
| 32 |
> High-precision **optical motion capture** data for humanoid robots, dexterous hands, embodied AI, and virtual production.
|
| 33 |
-
> 面向人形机器人、灵巧手、具身 AI 和虚拟制作的高精度多模态动作数据。
|
| 34 |
|
| 35 |
| | |
|
| 36 |
|---|---:|
|
|
@@ -42,20 +42,20 @@ dataset_info:
|
|
| 42 |
| **Formats / 格式** | BVH · Retargeted CSV · NPZ |
|
| 43 |
|
| 44 |
🔒 **Access note:** Metadata and samples are public. Full data requires **Request access**.
|
| 45 |
-
🔒 **访问说明:** 元数据和样本公开,完整数据需申请访问权限。
|
| 46 |
|
| 47 |
---
|
| 48 |
|
| 49 |
## Key Features / 核心特点
|
| 50 |
|
| 51 |
- **Optical ground truth** – sub-mm accuracy, 120 fps, no estimation errors.
|
| 52 |
-
**光学真值** – 亚毫米精度,120 fps,无
|
| 53 |
-
- **Dexterous hands** – 20+ DoF per hand, synchronized with object
|
| 54 |
-
**灵巧手数据** – 每只手 20+ 自由度,与物体
|
| 55 |
- **Robot-ready** – pre-retargeted to Unitree G1; custom retargeting available.
|
| 56 |
**机器人就绪** – 预重定向至 Unitree G1,可定制其他机器人。
|
| 57 |
-
- **Real-world diversity** – 15+ scenarios, 500+ tasks,
|
| 58 |
-
**真实场景多样性** – 15+ 场景,500+ 任务,
|
| 59 |
- **Multi-modal** – full-body skeleton, finger motion, object pose, multi-view video, semantic labels.
|
| 60 |
**多模态** – 全身骨骼、手指运动、物体位姿、多视角视频、语义标签。
|
| 61 |
- **Quality assured** – every take passes automated cleaning + manual inspection; quality flags provided.
|
|
|
|
| 27 |
---
|
| 28 |
|
| 29 |
# ChingMu 1000-Hour Embodied Motion Dataset
|
| 30 |
+
<small> 青瞳1000小时具身智能动作数据集</small>
|
| 31 |
|
| 32 |
> High-precision **optical motion capture** data for humanoid robots, dexterous hands, embodied AI, and virtual production.
|
| 33 |
+
><small> 面向人形机器人、灵巧手、具身 AI 和虚拟制作的高精度多模态动作数据。</small>
|
| 34 |
|
| 35 |
| | |
|
| 36 |
|---|---:|
|
|
|
|
| 42 |
| **Formats / 格式** | BVH · Retargeted CSV · NPZ |
|
| 43 |
|
| 44 |
🔒 **Access note:** Metadata and samples are public. Full data requires **Request access**.
|
| 45 |
+
🔒<small> **访问说明:** 元数据和样本公开,完整数据需申请**访问权限**。</small>
|
| 46 |
|
| 47 |
---
|
| 48 |
|
| 49 |
## Key Features / 核心特点
|
| 50 |
|
| 51 |
- **Optical ground truth** – sub-mm accuracy, 120 fps, no estimation errors.
|
| 52 |
+
**光学真值** – 亚毫米精度,120 fps,无累计误差。
|
| 53 |
+
- **Dexterous hands** – 20+ DoF per hand, synchronized with object 6DOP pose.
|
| 54 |
+
**灵巧手数据** – 每只手 20+ 自由度,与物体 6 DOP 同步。
|
| 55 |
- **Robot-ready** – pre-retargeted to Unitree G1; custom retargeting available.
|
| 56 |
**机器人就绪** – 预重定向至 Unitree G1,可定制其他机器人。
|
| 57 |
+
- **Real-world diversity** – 15+ scenarios, 500+ tasks, 200+ objects.
|
| 58 |
+
**真实场景多样性** – 15+ 场景,500+ 任务,200+ 物体。
|
| 59 |
- **Multi-modal** – full-body skeleton, finger motion, object pose, multi-view video, semantic labels.
|
| 60 |
**多模态** – 全身骨骼、手指运动、物体位姿、多视角视频、语义标签。
|
| 61 |
- **Quality assured** – every take passes automated cleaning + manual inspection; quality flags provided.
|