Archontis Politis

Generalizing DoA estimation across diverse microphone arrays could revolutionize audio processing in mobile and dynamic environments.

A training-free dynamic clustering method significantly improves long speech separation performance, especially in sparse scenarios with unknown speaker counts.

Achieving a mean angular error of just 11.3 degrees in head orientation estimation could revolutionize applications in smart environments and driver monitoring.

Upsampling covariance matrices with deep learning can transform a 4-microphone array into a powerful 32-microphone equivalent, revolutionizing acoustic imaging.

Perceptual tests show that a new transcoding framework significantly enhances spatial audio reproduction, especially for constrained microphone setups.

Relative cues like "who spoke first" or "who is louder" dramatically boost text-guided speech extraction, even outperforming systems that rely solely on audio.