I built a real-time collaborative CV annotation tool using CRDTs and WebRTC

Описание и применение

Инструмент Sightline CRDT позволяет командам машинного обучения совместно проверять и исправлять bounding‑boxes в реальном времени без конфликтов, используя Yjs CRDTs и WebRTC, с голосовым управлением и экспортом в COCO/YOLOv8.

Задача

Machine learning and computer vision teams need to audit and fix training bounding boxes collaboratively without merge conflicts.

Результат

The tool enables zero‑merge‑conflict collaborative annotation with voice control and instant export, improving audit efficiency.

Рабочий процесс

• Open the Sightline CRDT web app
• Select an image to annotate
• Invite teammates via the P2P link
• Use the UI or voice commands to adjust boxes
• Export annotations in COCO or YOLOv8 format

Доказательства

Sightline CRDT, an open in-browser tool for machine learning and computer vision teams to collaboratively audit and fix training bounding boxes without merge conflicts.

Источник

Powered by Yjs CRDTs over WebRTC so multiple auditors can inspect the same image simultaneously with zero merge conflicts.

Источник

Voice Patching: Speak verbal commands like "Nudge box left" to correct coordinates and labels hands‑free.

Источник

Full COCO and YOLOv8 export directly from browser memory.

Источник