• Intelligence Engineering Lab.

    Our laboratory researches Audio Processing with Deep Learning, Audio Manipulation with Textual Queries, Speech Enhancement/Denoising, Music Source Separation, and Generative Modeling.

Notice

SDX23 Challenge paper is now available on arxiv (2023)

We contributed on the paper "The Sound Demixing Challenge 2023 - Music Demixing Track". We collaborated with Sony, Meta AI, Tencent AI, and The University of Tokyo etc. Congrats to Minseok Kim and Jun Hyung Lee! [arxiv]

Notice

Our paper is now available on arxiv (2023)

Our paper 'Sound Demixing Challenge 2023 Music Demixing Track Technical Report: TFC-TDF-UNet v3' is now available on arxiv. We summarized our award-winning solutions for the Music Demxing Track of Sound Demixing Challenge 2023. [arxiv] [github] [more information]

Notice

We won 1st&3rd place in the mdx Challenge 2023

Music Demixing Track: Build A System To perform Audio Source Separation [more information]

Notice

Woosung Choi started working at Sony (2022)

Notice

Two silver medals at SONY MDX CHALLENGE @ ISMIR 2021

We proposed KUIELAB-MDX-Net, an hybrid architecture based on TFC-TDF-U-Net and Facebook’s Demucs, and we won two silver medals." [more information]

Notice

Our paper has been accepted to ACM MULTIMEDIA 2021

"AMSS-Net: Audio Manipulation on User-Specified Sources with Textual Queries" [demo] [github] [more information]

Notice

Our paper got accepted to ICASSP 2021

Our paper "LaSAFT: Latent Source Attentive Frequency Transformation for Conditioned Source Separation" got accepted to 2021 IEEE International Conference on Acoustics, Speech and Signal Processing! [demo] [github] [more information]

Ongoing Research Topic

Deep Learning-based Source-Selective Audio Manipulation Framework

“Ok google! 남자 목소리 줄여줘!” 같은 기능이 스마트폰에서 지원되면 얼마나 편할까요? 딥러닝 기반의 다중음원 오디오에 대한 음원선별적 편집 모델 연구에서는 여러 음원이 섞여 있는 signal에서 특정 음원만 선별적으로 편집하는 기법을 연구합니다.

Learn more

Ongoing Research Topic

Deep Learnig-based Blind Source Separation

현재 지능공학 연구실에서 수행중인 Deep Learnig-based Blind Source Separation 분야에 대한 소개글입니다.

Learn more

If you would like to get further information regarding our lab?

OUR PAPER GOT ACCEPTED TO ACM MULTIMEDIA 2021

- Woosung Choi, Minseok Kim, PH.D. Candidates

2021 한국연구재단 중견연구자지원사업 '딥러닝 기반의 다중음원 오디오에 대한 음원선별적 편집 모델 연구'선정

- 정순영, 지도교수

2020 한국연구재단 중견연구자지원사업 '다양한 음향신호처리에 범용적 적용이 가능한 신경망 기반 스펙트로그램 변환블록 연구' 선정

- 정순영, 지도교수