Automatic music feature extraction, classification and annotation

With the increasing amount of digital music being stored on digital devices and the Internet, techniques and systems are required to effectively and automatically process, classify and annotate music, so that relevant music can be located and retrieved quickly. This project will address fundamental research problems in automatic music feature extraction, classification and annotation by making use of psychoacoustic features and innovative classification techniques. The outcome will be useful in all areas where music is used including entertainment, digital library, the media, and Internet commerce.