---
configs:
- config_name: default
  data_files:
  - split: train
    path: viewer/train.csv
dataset_info:
  features:
  - name: package_id
    dtype: string
  - name: task_name
    dtype: string
  - name: task_description
    dtype: string
  - name: modalities
    dtype: string
  - name: package_path
    dtype: string
  - name: training_ready_path
    dtype: string
  - name: package_size_gb
    dtype: float64
  - name: validation
    dtype: string
  splits:
  - name: train
    num_examples: 8
license: cc-by-nc-4.0
task_categories:
- robotics
- depth-estimation
- image-segmentation
- keypoint-detection
- time-series-forecasting
tags:
- egocentric
- rgb-d
- emg
- imu
- hand-object-contact
- per-finger-force
- robotics
pretty_name: Egocentric RGB-D + EMG/IMU Daily Activity Dataset
---
# Egocentric RGB-D + EMG/IMU Daily Activity Dataset

This dataset contains first-person daily activity recordings with synchronized RGB-D video, wrist EMG/IMU signals, hand keypoints, object masks, hand-object contact annotations, per-finger force annotations, and semantic action segments.

![RGB-D and EMG multimodal examples](examples/rgb_depth_emg_multimodal_grid.jpg)

**Multimodal showcase video:** [RGB-D, hand joints, 3D hand projection, and object masks](examples/rgb_depth_hand_mask_showcase.mp4)

## Overview

The dataset is designed for egocentric embodied AI and robot learning in everyday household activities. RGB-D frames provide visual and geometric context, EMG/IMU streams capture hand activity, and contact-force annotations describe where fingers interact with objects and how force is distributed across the hand.

These modalities support research on manipulation understanding, imitation learning, action segmentation, contact-aware perception, multimodal behavior modeling, and robot training from human demonstrations.

## Dataset Highlights

- First-person RGB-D recordings from daily household tasks.
- Synchronized wrist EMG/IMU streams aligned to the RGB frame timeline.
- Hand keypoints, object masks, hand-object contact annotations, and per-finger force annotations.
- Per-frame 21-point hand keypoint JSONL exports are included under `analysis/hand_keypoints/`.
- Public object-mask preview videos are included for each package under `analysis/object_mask_public/`.
- Semantic action segments for task-level and subtask-level analysis.
- `training_ready/` exports for model and robot learning workflows.

## Packages

| Package | Files | Size GB | Validation |
|---|---:|---:|---|
| `washing-machine-laundry` | 334 | 1.503 | pass |
| `tidy-bedroom` | 1246 | 6.076 | pass |
| `sweep-and-mop-floor` | 1395 | 7.163 | pass |
| `tidy-living-room` | 935 | 4.132 | pass |
| `tidy-dining-room` | 1616 | 7.711 | pass |
| `hand-cream-application` | 263 | 0.551 | pass |
| `sink-hand-washing` | 157 | 0.353 | pass |
| `paper-towel-hand-wiping` | 165 | 0.383 | pass |

## Structure

```text
packages/<package_id>/
  source_stage_a/          # original lossless RGB-D and sensor export
  clean/                   # aligned timelines and corrected sensors
  gold/                    # episode and standard model exports
  analysis/contact_force_v2/
  analysis/hand_keypoints/
  analysis/object_mask_public/
  analysis/semantic_subtasks/
  events/
  training_ready/          # consolidated model-ready files
```

## Training

For model or robot training, start from `packages/<package_id>/training_ready/`. It contains frame indexes, corrected EMG/IMU streams, semantic segments, contact-force tables, and standard imitation-learning exports when available.

## Loader Examples

The repository includes two lightweight Python examples under `examples/`.

```bash
python examples/quickstart_loader.py --dataset-root .
python examples/training_sample_loader.py --dataset-root . --package-id hand-cream-application
```

`quickstart_loader.py` reads `viewer/train.csv`, lists all packages, and checks the expected public directories. `training_sample_loader.py` shows how to locate RGB/depth timelines, corrected EMG/IMU streams, semantic action segments, contact-force tables, RLDS episodes, and robomimic exports for one package.

The quickstart script uses only the Python standard library. The training sample script can also preview Parquet tables when `pandas` and `pyarrow` are installed.
## Citation

If you use this dataset, please cite the dataset repository and reference the package IDs used in your experiments.