Generalized Recognition of Sound Events: Approaches and Applications

This chapter surveys the contemporary approaches of automatic sound recognition and discusses the benefits stemming from real-world applications of this technology. We identify the common aspects and subtle differences among these diverse application areas and review state-of-the-art systems. In this context we project that there is much space for knowledge transfer between the different subfields of sound classification, which seem to evolve independently while achieving different states of maturity. Particular emphasis is given to lessons learned from the speech recognition paradigm, which together with speaker recognition were among the first applications of sound classification that reached the status of launching commercial products at a large climax. Special attention is paid to new emerging applications such as environmental monitoring and bioacoustic identification and applications to music which have already started altering our everyday life as we once knew it.