Signal Processing and Communications Laboratory

Department of Engineering

James Cozens James Cozens

Background - Research

Position: PhD Student

E-mail: jmc257 [at] cam.ac.uk

Office Location: BN3-02

Thesis Title: Architecting Expressional Landscapes: Hierarchical Generative Models for Artistic Media

Supervisor: Prof. Simon Godsill

Background

Research Interests

His research focuses on statistical signal processing and machine learning methods in the context of audio and music processing, including high-resolution time-frequency analysis, music transcription, beat tracking, and signal decomposition. Additional interests involve multi-object tracking and deep learning-based hierarchical generative models for music visualisation and composition, with an emphasis on achieving expressive and structural consistency.