Detector of blogger advertisements in YouTube videos
Goal of project: to develop Web browser extension to detect and automatically skip advertisements in YouTube videos which where inserted by blogger at video editing stage.
Components:
- AdMark - web application to markup advertisements in videos to collect training data.
- AdDetectorModel - consists of main logic for detecting ads.
- AdDetectorChromeExtension - browser extension for Google Chrome that requests location of ads from server by video identifier and automatically skips these parts of video.
- AdDetectorServer - REST server that responds to queries from browser extension and translate queries to workers via message queue.
- AdDetectorWorker - handler of queries, it gets queries from server via message queue, uses model to detect ads and saves results to database.