ASPIRE

Audio and Signal Processing, Interpreration and Enhancement

The Audio & Signal Processing, Interpretation and Enhancement (ASPIRE@IDLab) group is the research group headed by Prof. Dr.-Ing Nilesh Madhu at Ghent University. Our goal is the robust capture, enhancement, and analysis of signals, for information extraction in a wide variety of applications.

We have three main research pillars:

• Audio and speech processing where we have state-of-the-art solutions for
  • localising and segregating sources for telecommunications and smart-home applications; and
  • optimal audio reproduction in extended/virtual reality
• Signal processing for biomedical applications, with ongoing research into
  • nanopore analysis for sequencing
  • neural signal compression and transmission
• Representation learning, focussing on optimal extraction and encoding of relevant information in latent spaces. This is a fundamental topic, with wide ranging applications such as scene analysis, emotion and stress recognition, ...

The underlying philosophy in our approaches is the appropriate embedding of domain-specific knowledge within deep-learning frameworks. This can better leverage the available data, while allowing the networks to generalise well to unseen conditions in real-life situations.

Demos and more details of our work are available at aspire.ugent.be