---
license: cc-by-nc-4.0
task_categories:
  - robotics
  - depth-estimation
tags:
  - tracegen
  - traceforge
  - epic-kitchens
  - computer-vision
  - non-commercial
  - 3d-trajectories
---

# TraceGen – EPIC-KITCHENS (Derived Subset)

## Overview
This folder contains a **derived subset** of the EPIC-KITCHENS dataset processed using the
TraceForge pipeline as part of the TraceGen project.

- **TraceGen Project Website**: https://tracegen.github.io/

This data is released **strictly for non-commercial academic research purposes**.

## Evaluation Protocol

This dataset defines the official evaluation protocol for the TraceGen benchmark.
Models are evaluated on five environments with the following metrics:
- Mean Squared Error (MSE)
- Mean Absolute Error (MAE)
- Endpoint MSE

The official datasets and models are hosted at:
👉 [TraceGen Collection](https://huggingface.co/collections/furonghuang-lab/tracegen)

The official leaderboard is hosted at:
👉 [TraceGen Leaderboard](https://huggingface.co/furonghuang-lab/TraceGenLeaderboard)

## Source Dataset
- **Dataset**: EPIC-KITCHENS
- **Original Authors**: EPIC-KITCHENS Consortium
- **Original License**: Creative Commons Attribution-NonCommercial 4.0 (CC BY-NC 4.0)
- **Official Website**: https://epic-kitchens.github.io/

## What Is Released Here
This folder contains a subset of episodes processed by the TraceForge pipeline.
For each episode `<video_name>`, we release:

- `images/`: RGB frames extracted from the original EPIC-KITCHENS videos (`*.png`)
- `depth/`: predicted depth maps aligned to the released frames (`*.png`, `*_raw.npz`)
- `samples/`: per-frame TraceForge / TraceGen representations (`*.npz`)
- `<video_name>.npz`: aggregated episode-level metadata and outputs

### Notes
- RGB frames originate from the EPIC-KITCHENS dataset.
- Depth and trace representations are **derived outputs** generated by TraceForge.

## Modifications
Processing steps include:
- frame subsampling and temporal alignment
- camera pose estimation and compensation
- depth prediction and 3D reconstruction
- trace abstraction and normalization

## License and Attribution
This derived dataset remains subject to the **CC BY-NC 4.0** license.

Users must:
- attribute the EPIC-KITCHENS dataset and authors
- clearly indicate that this is a derived dataset
- **not use this data for commercial purposes**

## Disclaimer
This release **does not replace, mirror, or redistribute the original EPIC-KITCHENS dataset**.
All rights to the original data remain with the EPIC-KITCHENS authors.