Welcome to IndEgo, a NeurIPS 2025 Datasets & Benchmarks Track accepted dataset and open-source framework for industrial egocentric vision, designed to support training, real-time guidance, process improvement, and collaboration.
🎓 Paper: IndEgo: A Dataset of Industrial Scenarios and Collaborative Work for Egocentric Assistants (NeurIPS 2025)
👉 GitHub Repo
🤗 Hugging Face Dataset
🚀 
🎥 Industrial Scenarios
📘 About
IndEgo introduces a multimodal egocentric + exocentric video dataset capturing common industrial activities such as assembly/disassembly, inspection, repair, logistics, and woodworking.
It includes 3,460 egocentric videos (~197h) and 1,092 exocentric videos (~97h) with synchronised eye gaze, audio narration, hand pose, motion, and semi-dense point clouds.
IndEgo enables research on:
- Procedural & collaborative task understanding
- Mistake detection and process deviation recognition
- Reasoning-based Video Question Answering (VQA)
⚙️ Technology
IndEgo combines:
- Egocentric Computer Vision for context-aware task understanding
- Vision-Language Models (VLMs) for multimodal reasoning
- Smart Glasses Integration for on-site, real-time assistance
🎬 Demo Video
🚀 Try It: No Setup Required
Launch Colab Notebook
Run IndEgo’s core logic directly in your browser with Google Colab — no installation needed.
📊 Dataset
🔗 Open Dataset on Hugging Face
The IndEgo dataset includes annotated egocentric and exocentric videos of real-world industrial scenarios with:
- Action & narration annotations
- Mistake labels and summaries
- Eye-gaze and 3D mapping data
- Benchmarks for procedural reasoning and collaborative task understanding
🧩 Citation
If you use IndEgo, please cite:
bibtex
@inproceedings{Chavan2025IndEgo,
author = {Vivek Chavan and Yasmina Imgrund and Tung Dao and Sanwantri Bai and Bosong Wang and Ze Lu and Oliver Heimann and J{\"o}rg Kr{\"u}ger},
title = {IndEgo: A Dataset of Industrial Scenarios and Collaborative Work for Egocentric Assistants},
booktitle = {Advances in Neural Information Processing Systems (NeurIPS) Datasets and Benchmarks Track},
year = {2025},
url = {https://neurips.cc/virtual/2025/poster/121501}
}