Semantically Meaningful Music Retrieval with Content-based Features and Fuzzy Clustering

We propose a system which is based on fuzzy c-means clustering and automatically organizes a collection of music files according to musical surface characteristics (e.g., zero crossings, short-time energy, etc.) and tempo. Our system is complemented by semantic metadata which offers the user the ability to check the clustering result, correct existing textual meta-information (such as genre and album) and add additional semantic information (about lyrics, theme, other artists who perform similar music, musical instruments, etc). This second step is realized in a process of relevance feedback and imposes consistency and reliability between content and semantic information of the musical database.