Skip to content
Max Hild edited this page Feb 17, 2026 · 1 revision

Welcome to the lx-Annotate wiki!

Lx-Annotate is a comprehensive study setup for annotating videos and creating medical reports. The software was developed during the ColoReg study to provide a frontend for a reusable register study.

This is why the dependencies include endoreg-db, a postgres database.

Lx-Annotate works best in production on LuxNix machines, where it is available as a system service. Otherwise it needs to be run in development mode or using the (experimental) Docker container.

The state of Lx-Annotate depends on the production environment. Secretspec.toml provides defaults for the settings, production and development modes will change a number of security settings and dev will enable debug mode for easier modification.

The Data import is handled by a automated file watcher. The system service will start automatically when lx-Annotate is run in a development environment. Otherwise, it is started using a external service that will run a file watcher instance. File watcher starts ingesting the video, which means applying the default anonymization pipeline. Depending on the environment, OCR NER and LLM Processing is done to extract the text of an image. The default settings apply, while they are changeable.

Clone this wiki locally