NMF/NTF-based methods applied for user-guided audio source separation: An overview
Manh-Quan Bui, Viet-Hang Duong, Shih-Pang Tseng, Zhao-Ze Hong, Bo-Chang Chen, Zhi-Wei Zhong, Jia‐Ching Wang · 2016
Recently, user-guided or user-assisted audio separation, which exploits the auxiliary information of sources provided by users, is a new direction in audio source separation. Nonnegative matrix or tensor factorization (NMF/NTF) is one of the promising classes of approaches to address this task. In this paper, we present a survey of recent developments in user-guided source separation approaches based on NMF/NTF and illustrate various ways for incorporating the provided knowledge from user.