Development and Optimization of Music Processing Algorithms for Cochlear Implant Users

Practical project, Bachelor's thesis & Master's thesis

Topic Description
Cochlear implants (CI) can make hearing possible again for people with profound hearing loss or deafness. The auditory nerve is electrically stimulated by an electrode array implanted in the cochlea. In many cases, this can restore particularly speech comprehension in particular; whereas music is perceived as too complex and therefore often unpleasant, especially with high polyphony. To reduce the complexity, various music processing algorithms have been developed in recent years. Auditory models that simulate the reaction of the auditory nerve to input audio signals can be used to validate and further develop these algorithms.

Task Description
Auditory models are to be used for the development and validation of new music processing methods. One goal is to develop efficient DNN-based approximations of current auditory models (for normal hearing and CI users) that accelerate the simulation of neural responses. For the optimization of music processing algorithms, objective evaluation criteria such as the Neural Similarity Measure (NSIM) will be further developed to estimate the similarity of simulated neural responses in normal hearing and CI users.

Required Background

• Interest in music processing and cochlear implant technology
• Advanced understanding of systems and signals
• Good programming skills