Detection of Flashy Segment in YouTube Videos using Computer Vision Technique to Prevent Seizure in Epilepsy Patients

Epilepsy is a medical illness that affects roughly 50 million people worldwide, with 3-5% of these people experiencing seizures caused by flashing or flickering lights or specific patterns with areas of light and dark contrast. This condition is referred to as photosensitive epilepsy (PSE). YouTube is a leading video-sharing website that contains video content that often has rapid changes in brightness or contains strong contrasts in spatial patterns that can trigger seizures. It is important to detect such content before it is distributed or displayed to prevent adverse effects on vulnerable individuals. The proposed solution is a browser extension that provides real-time scanning and processing of YouTube videos to detect and disarm potential seizure-inducing segments. The extension is compatible with Chromium-related browsers, including Brave, and is designed to display a warning if epileptic content is present in a YouTube video. The system utilizes computer vision techniques such as OpenCV to detect flashing video content that complies with the CIE 1931 color space, enabling it to skip seizure-triggering segments from YouTube videos.