Release of DeFT-AN (M-ch Speech Enhancement)

Dongheon Lee, a Ph.D. student, showcased his second multichannel speech enhancement model (DeFT-AN) at the ICASSP 2023 conference in Greece. Congratulations to him on his inaugural ICASSP presentation! We also wish him good luck with the upcoming release of the real-time version (DeFT-AN RT) at InterSpeech 2023.

For access to the codes and sample speech clips, please visit https://github.com/donghoney0416/DeFT-AN.

The complete list of our datasets & codes can be found at https://www.sound.kaist.ac.kr/opensrcs

The paper has been published in IEEE SPL, and you can find it at the following link: https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=10042963.