Research Areas
Speech Enhancement / Separation
Aims to improve speech clarity in non-studio environments where microphone or environmental noise causes signal degradation.
Binaural Rendering
Utilizes environmental context and DNN architectures to regenerate immersive 3D audio scenes optimized for specific playback spaces.
Room Geometry Inference
Estimates complex indoor geometries using acoustic echoes to discover hidden areas and create accurate digital twins.
Anomaly Detection
Identifies atypical acoustic patterns to detect equipment failure or environmental disturbances using self-supervised machine learning models.
Target Sound Extraction
Isolates desired signals from noisy mixtures using Transformer-based models guided by class labels and spatial direction clues.
Sound Event Localization & Detection
Classifies and localizes 3D sound events using independent attention mechanisms applied to spectral and spatial acoustic domains.
News
+Media Release
+Recent Publications
+- Distilling Selective Patches for Speech EnhancementDoHwan Kim, Jung-Woo Choi•IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Barcelona, Spain, 2026•2026
- Can Large Audio Language Models Understand Audio Well?Han Yin, Jung-Woo Choi•IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Barcelona, Spain, 2026•2026
- Navigating Target Sound Extraction With Effective Directional Clue IntegrationDayun Choi, Jung-Woo Choi•IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Barcelona, Spain, 2026•2026