An interactive, browser-based simulator that shows how changes in log intensity become positive and negative events.
- Converts a live webcam feed or local video into simulated event-camera data.
- Visualizes positive and negative events with adjustable persistence.
- Lets you inspect any pixel's intensity, log intensity, reference level, and threshold crossings.
- Includes ideal and simplified realistic sensor modes, asymmetric thresholds, preprocessing controls, and guided experiments.
- Explains the limitations of generating events from conventional frame-based video.
All camera and video processing happens locally in the browser. Media is never uploaded, recorded, or sent to a backend.
Requires Node.js 22.13 or newer.
npm install
npm run devThen open http://localhost:3000.
npm run build
npm run lintFrames are resized, converted to grayscale, and processed in a Web Worker. For each pixel, the simulator compares the current log intensity with the reference stored at its previous event. Threshold crossings emit one or more events and advance that reference by the corresponding contrast threshold.
Because a webcam only supplies frames, event timestamps between frames are linearly estimated. This is an educational simulator, not a replacement for a hardware event camera with asynchronous microsecond timing and high dynamic range.
React, TypeScript, vinext, Canvas, Web Workers, and OpenAI Sites.
- Event-based Vision: A Survey
- v2e: From Video Frames to Realistic DVS Events
- The Event-Camera Dataset and Simulator
No license has been added yet.
