IEEE ICASSP 2022

2022 IEEE International Conference on Acoustics, Speech and Signal Processing

7-13 May 2022
  • Virtual (all paper presentations)
22-27 May 2022
  • Main Venue: Marina Bay Sands Expo & Convention Center, Singapore
27-28 October 2022
  • Satellite Venue: Crowne Plaza Shenzhen Longgang City Centre, Shenzhen, China

ICASSP 2022
SPE-33.5

A TRAINING FRAMEWORK FOR STEREO-AWARE SPEECH ENHANCEMENT USING DEEP NEURAL NETWORKS

Bahareh Tolooshams, Harvard University, United States of America; Kazuhito Koishida, Microsoft Corporation, United States of America

Session:
Speech Enhancement: Training Schemes and Losses

Track:
Speech and Language Processing

Location:
Gather Area B

Presentation Time:
Tue, 10 May, 20:00 - 20:45 China Time (UTC +8)
Tue, 10 May, 12:00 - 12:45 UTC

Session Chair:
Kazuhito Koishida, Microsoft Corporation
Presentation
Discussion
Resources
Session SPE-33
SPE-33.1: PHASE CONTINUITY: LEARNING DERIVATIVES OF PHASE SPECTRUM FOR SPEECH ENHANCEMENT
Doyeon Kim, Hyewon Han, Hong-Goo Kang, Yonsei University, Korea, Republic of; Hyeon-Kyeong Shin, Soo-Whan Chung, Naver Corporation, Korea, Republic of
SPE-33.2: CONTINUAL SELF-TRAINING WITH BOOTSTRAPPED REMIXING FOR SPEECH ENHANCEMENT
Efthymios Tzinis, University of Illinois at Urbana-Champaign, United States of America; Yossi Adi, Meta AI Research, Israel; Vamsi Ithapu, Buye Xu, Anurag Kumar, Meta Reality Labs Research, United States of America
SPE-33.3: Alleviating the Loss-Metric Mismatch in Supervised Single-Channel Speech Enhancement
Yang Yang, Hui Zhang, Xueliang Zhang, Huaiwen Zhang, Inner Mongolia University, China
SPE-33.4: A PRIORI SNR ESTIMATION FOR SPEECH ENHANCEMENT BASED ON PESQ-INDUCED REINFORCEMENT LEARNING
Tong Lei, Haoxin Ruan, Kai Chen, Jing Lu, Key Laboratory of Modern Acoustics, Nanjing University; NJU-Horizon Intelligent Audio Lab, Horizon Robotics; Nanjing Institute of Advanced Artificial Intelligence., China
SPE-33.5: A TRAINING FRAMEWORK FOR STEREO-AWARE SPEECH ENHANCEMENT USING DEEP NEURAL NETWORKS
Bahareh Tolooshams, Harvard University, United States of America; Kazuhito Koishida, Microsoft Corporation, United States of America
SPE-33.6: Joint magnitude estimation and phase recovery using Cycle-in-Cycle GAN for non-parallel speech enhancement
Guochen Yu, Communication University of China/Institute of Acoustics, Chinese Academy of Sciences, China; Andong Li, Chengshi Zheng, Institute of Acoustics, Chinese Academy of Sciences, China; Yutian Wang, Hui Wang, Communication University of China, China; Yinuo Guo, Bytedance, China