Instrument for Soundscape Recognition, Identification and Evaluation (ISRIE): Signal Classification

ISRIE is a collaborative project between the universities of York and Newcastle and ISVR in Southampton. The work being undertaken at York is in its second year and focuses on signal separation and classification. Developing novel methods for classifying urban and other sounds into distinct categories (such as transportation, industrial, human, animal, etc.) is the focus of this project. The classification system will consist of two main parts: a feature extractor and a classifier. Both time and time‐frequency domain techniques have been considered for feature extraction and it is anticipated that a combination of domains will be required for this complex task. To date only simple classifiers have been implemented and development is now focussed on syntactic pattern recognition. The project background and an overview of the techniques involved will be presented along with preliminary results for the classification of urban and other sounds using syntactic methods.