Documentation

CV-Scope runs locally without any cloud service. The backend is a Python FastAPI application; the frontend is a React application served by Vite in development or bundled into the backend in…

The Windows installer sets up everything CV-Scope needs in one go: a private copy of Python, PyTorch for your processor or NVIDIA graphics card, the detection models you pick and the sample videos.…

This guide walks through the first experiment: a route-choice study on a recorded video. The same steps apply to a live camera and to vehicles.

CV-Scope does not hard-wire one detector. The model catalog (backend/pathscope/models/catalog.py) lists every supported model with its family, task, classes, size, approximate memory and compute…

The Anomaly Assistant reports meaningful visual change from an expected, static scene: objects that appear, disappear or move, presence where an area should be empty, a disturbed or rearranged scene,…

CV-Scope's relationship engine connects what the cameras observe into time-bounded, confidence-scored relationships, and correlates them into higher-level events:

The relationship engine (docs/relationships.md) works inside one camera's run. This layer places cameras in the real world and connects what several cameras saw:

CV-Scope's open-source core tracks people and vehicles anonymously. Two optional, proprietary modules extend it for authorised deployments:

A worked example of locations, cross-camera moves, entities and the relationship graph, built from real multi-camera footage. It lives in the project Sample - MEVA school campus and the location…

CV-Scope is built to answer questions about movement without identifying anyone. The defaults below hold for every installation; the Settings page shows the same list live for the current…

The database (SQLite by default, PostgreSQL optional) separates configuration, execution, research records, validation and platform state. Raw frames are never stored.

CV-Scope is a configurable computer-vision monitoring and research platform for fixed cameras. This document records the architecture assessment, the technology decisions, the folder layout, the…

Each run is one worker process (multiprocessing, spawn context) started by RunSupervisor. The worker builds a Pipeline, loops over frames and sends:

Thank you for considering a contribution. CV-Scope is a research and monitoring platform; changes that make it more truthful, more robust or easier to use for non-programmers are the most valuable…

Please report security problems privately, not in a public issue: use GitHub's Report a vulnerability button on the repository's Security tab (private vulnerability reporting). Include the version or…

How the repository is protected against bad or malicious changes, the GitHub settings that go with the files in .github/, and how to review a pull request safely.