Update README with motion annotations docs, usage examples, and known limitations

README.md

@@ -10,16 +10,20 @@ tags:
   - manipulation
   - 6dof
   - mediapipe
+  - egocentric
+  - motion-semantics
 size_categories:
   - 10K<n<100K
 language:
   - en
-pretty_name: Dynamic Intelligence Humanoid Robots Training Dataset
+pretty_name: Dynamic Intelligence - Egocentric Human Motion Annotation Dataset
 ---
 
-# Dynamic Intelligence - Humanoid Robots Training Dataset
+# Dynamic Intelligence - Egocentric Human Motion Annotation Dataset
 
-RGB-D hand manipulation dataset captured with iPhone 13 TrueDepth sensor for humanoid robot training.
+RGB-D hand manipulation dataset captured with iPhone 13 TrueDepth sensor for humanoid robot training. Includes 6-DoF hand pose trajectories, synchronized video, and semantic motion annotations.
+
+---
 
 ## 📊 Dataset Overview
 
@@ -28,47 +32,130 @@ RGB-D hand manipulation dataset captured with iPhone 13 TrueDepth sensor for hum
 | Episodes | 97 |
 | Total Frames | ~28,000 |
 | FPS | 30 |
-| Tasks | 10 manipulation tasks (folding, picking, placing) |
+| Tasks | 10 manipulation tasks |
+| Total Duration | ~15.5 minutes |
+| Avg Episode Length | ~9.6 seconds |
+
+### Task Distribution
+
+| Task ID | Description | Episodes |
+|---------|-------------|----------|
+| Task 1 | Fold the white t-shirt on the bed | 8 |
+| Task 2 | Fold the jeans on the bed | 10 |
+| Task 3 | Fold two underwear and stack them | 10 |
+| Task 4 | Put the pillow on the right place | 10 |
+| Task 5 | Pick up plate and glass, put on stove | 10 |
+| Task 6 | Go out the door and close it | 9 |
+| Task 7 | Pick up sandals, put next to scale | 10 |
+| Task 8 | Put cloth in basket, close drawer | 10 |
+| Task 9 | Screw the cap on your bottle | 10 |
+| Task 10 | Pick up two objects, put on bed | 10 |
 
 ---
 
-## 📁 Structure
+## 📁 Repository Structure
 
 ```
-data/
-├── data/chunk-000/          # Parquet files per episode
-│   ├── episode_000000.parquet
-│   ├── episode_000001.parquet
-│   └── ...
-├── videos/chunk-000/rgb/    # MP4 videos per episode
-│   ├── episode_000000.mp4
-│   └── ...
-└── meta/
-    ├── info.json            # Dataset metadata
-    ├── stats.json           # Feature statistics
-    ├── events.json          # Disturbance/recovery annotations
-    └── depth_quality_summary.json  # Depth QC metrics
+humanoid-robots-training-dataset/
+│
+├── data/
+│   └── chunk-000/                    # Parquet files (97 episodes)
+│       ├── episode_000000.parquet
+│       ├── episode_000001.parquet
+│       └── ...
+│
+├── videos/
+│   └── chunk-000/rgb/                # MP4 videos (synchronized)
+│       ├── episode_000000.mp4
+│       └── ...
+│
+├── meta/                             # Metadata & Annotations
+│   ├── info.json                     # Dataset configuration (LeRobot format)
+│   ├── stats.json                    # Feature min/max/mean/std statistics
+│   ├── events.json                   # Disturbance & recovery annotations
+│   ├── depth_quality_summary.json    # Per-episode depth QC metrics
+│   └── annotations_motion_v1_frames.json  # Motion semantics annotations
+│
+└── README.md
 ```
 
 ---
 
-## 🎯 Parquet Schema (per episode)
+## 🎯 Data Schema
+
+### Parquet Columns (per frame)
 
 | Column | Type | Description |
 |--------|------|-------------|
 | `episode_index` | int64 | Episode number (0-96) |
 | `frame_index` | int64 | Frame within episode |
 | `timestamp` | float64 | Time in seconds |
-| `language_instruction` | string | Task description (e.g., "Fold the white t-shirt") |
-| `observation.state` | float[90] | 15 hand joints × 6 DoF (x,y,z in cm + yaw,pitch,roll in deg) |
-| `action` | float[90] | Same as state (for imitation learning) |
+| `language_instruction` | string | Task description |
+| `observation.state` | float[252] | 21 hand joints × 2 hands × 6 DoF |
+| `action` | float[252] | Same as state (for imitation learning) |
 | `observation.images.rgb` | struct | Video path + timestamp |
 
+### 6-DoF Hand Pose Format
+
+Each joint has 6 values: `[x_cm, y_cm, z_cm, yaw_deg, pitch_deg, roll_deg]`
+
+**Coordinate System:**
+- Origin: Camera (iPhone TrueDepth)
+- X: Right (positive)
+- Y: Down (positive)  
+- Z: Forward (positive, into scene)
+
 ---
 
-## 📋 Events Metadata (`meta/events.json`)
+## 🏷️ Motion Semantics Annotations
+
+**File:** `meta/annotations_motion_v1_frames.json`
 
-Annotated disturbances and recovery actions for select episodes.
+Coarse temporal segmentation with motion intent, phase, and error labels.
+
+### Annotation Schema
+
+```json
+{
+  "episode_id": "Task1_Vid2",
+  "segments": [
+    {
+      "start_frame": 54,
+      "end_frame_exclusive": 140,
+      "motion_type": "grasp",           // What action is being performed
+      "temporal_phase": "start",        // start | contact | manipulate | end
+      "actor": "both_hands",            // left_hand | right_hand | both_hands
+      "target": {
+        "type": "cloth_region",         // cloth_region | object | surface
+        "value": "bottom_edge"          // Specific target identifier
+      },
+      "state": {
+        "stage": "unfolded",            // Task-specific state
+        "flatness": "wrinkled",         // For folding tasks only
+        "symmetry": "asymmetric"        // For folding tasks only
+      },
+      "error": "none"                   // misalignment | slip | drop | none
+    }
+  ]
+}
+```
+
+### Motion Types
+`grasp` | `pull` | `align` | `fold` | `smooth` | `insert` | `rotate` | `open` | `close` | `press` | `hold` | `release` | `place`
+
+### Why Motion Annotations?
+- **Temporal Structure**: Know when manipulation phases begin/end
+- **Intent Understanding**: What the human intends to do, not just kinematics
+- **Error Detection**: Labeled failure modes (slip, drop, misalignment)
+- **Training Signal**: Richer supervision for imitation learning
+
+---
+
+## 📋 Events Metadata
+
+**File:** `meta/events.json`
+
+Disturbances and recovery actions for select episodes.
 
 ### Disturbance Types
 | Type | Description |
@@ -89,29 +176,33 @@ Annotated disturbances and recovery actions for select episodes.
 
 ---
 
-## 📈 Depth Quality Metrics (`meta/depth_quality_summary.json`)
+## 📈 Depth Quality Metrics
 
-Quality control metrics for each episode's depth data.
+**File:** `meta/depth_quality_summary.json`
 
 | Metric | Description | Dataset Average |
 |--------|-------------|-----------------|
-| `valid_depth_pct` | % of frames with valid depth at hand location | 95.5% ✅ |
-| `plane_rms_mm` | RMS deviation from flat surface (mm) | 5.73mm ✅ |
+| `valid_depth_pct` | % frames with valid depth at hand | 95.5% ✅ |
+| `plane_rms_mm` | RMS deviation from flat surface | 5.73mm ✅ |
 
 ---
 
 ## 🔧 Capture Setup
 
-- **Device:** iPhone 13 Pro (TrueDepth front camera)
-- **RGB Resolution:** 640×480 @ 30fps
-- **Depth Resolution:** 640×480 (synchronized with RGB)
-- **Depth Format:** Meters (float32)
-- **Hand Tracking:** MediaPipe (21 landmarks → 15 selected joints)
-- **6-DoF Computation:** Back-projection + vector kinematics
+| Parameter | Value |
+|-----------|-------|
+| **Device** | iPhone 13 Pro (TrueDepth front camera) |
+| **RGB Resolution** | 640×480 @ 30fps |
+| **Depth Resolution** | 640×480 (synchronized) |
+| **Depth Format** | Meters (float32) |
+| **Hand Tracking** | MediaPipe (21 landmarks per hand) |
+| **6-DoF Computation** | Back-projection + vector kinematics |
 
 ---
 
-## 🚀 Usage with LeRobot
+## 🚀 Usage
+
+### With LeRobot
 
 ```python
 from lerobot.common.datasets.lerobot_dataset import LeRobotDataset
@@ -120,13 +211,80 @@ dataset = LeRobotDataset("DynamicIntelligence/humanoid-robots-training-dataset")
 
 # Access episode
 episode = dataset[0]
-state = episode["observation.state"]  # [90] hand pose
-rgb = episode["observation.images.rgb"]  # Video frame
+state = episode["observation.state"]      # [252] hand pose (both hands)
+rgb = episode["observation.images.rgb"]   # Video frame
+task = episode["language_instruction"]    # Task description
+```
+
+### Loading Motion Annotations
+
+```python
+import json
+from huggingface_hub import hf_hub_download
+
+# Download annotations
+path = hf_hub_download(
+    repo_id="DynamicIntelligence/humanoid-robots-training-dataset",
+    filename="meta/annotations_motion_v1_frames.json",
+    repo_type="dataset"
+)
+
+with open(path) as f:
+    annotations = json.load(f)
+
+# Get segments for Task1
+task1_episodes = annotations["tasks"]["Task1"]["episodes"]
+for ep in task1_episodes:
+    print(f"{ep['episode_id']}: {len(ep['segments'])} segments")
+```
+
+### Combining Pose + Annotations
+
+```python
+# Get frame-level motion labels
+def get_motion_label(frame_idx, segments):
+    for seg in segments:
+        if seg["start_frame"] <= frame_idx < seg["end_frame_exclusive"]:
+            return seg["motion_type"], seg["temporal_phase"]
+    return None, None
+
+# Example: label each frame
+for frame_idx in range(episode["frame_index"].max()):
+    motion, phase = get_motion_label(frame_idx, episode_annotations["segments"])
+    if motion:
+        print(f"Frame {frame_idx}: {motion} ({phase})")
+```
+
+---
+
+## ⚠️ Known Limitations
+
+1. **Depth Dropouts**: Some frames have invalid depth in low-light or high-motion scenarios (~4.5% of frames)
+2. **Hand Tracking Accuracy**: MediaPipe may lose tracking during fast movements or self-occlusion
+3. **Single Viewpoint**: Egocentric only - no multi-view coverage
+4. **Limited Object Diversity**: Same objects used across episodes within each task
+5. **Annotation Coverage**: Motion annotations cover primary manipulation phases; transitional frames may be unlabeled
+
+---
+
+## 📖 Citation
+
+If you use this dataset in your research, please cite:
+
+```bibtex
+@dataset{dynamic_intelligence_2024,
+  author = {Dynamic Intelligence},
+  title = {Egocentric Human Motion Annotation Dataset},
+  year = {2024},
+  publisher = {Hugging Face},
+  url = {https://huggingface.co/datasets/DynamicIntelligence/humanoid-robots-training-dataset}
+}
 ```
 
 ---
 
 ## 📧 Contact
 
-**Organization:** [Dynamic Intelligence](https://huggingface.co/DynamicIntelligence)
+**Organization:** [Dynamic Intelligence](https://dynamicintelligence.company)
 
+For questions or collaboration inquiries, please open a Discussion on this repository.