Multi-Scale multi-band densenets for audio source separation (2017-06-29T00:00:00.000000Z)

TL;DR

A novel network architecture that extends the recently developed densely connected convolutional network (DenseNet) and takes advantage of long contextual information and outperforms state-of-the-art results on SiSEC 2016 competition by a large margin in terms of signal-to-distortion ratio.

Abstract

This paper deals with the problem of audio source separation. To handle the complex and ill-posed nature of the problems of audio source separation, the current state-of-the-art approaches employ deep neural networks to obtain instrumental spectra from a mixture. In this study, we propose a novel network architecture that extends the recently developed densely connected convolutional network (DenseNet), which has shown excellent results on image classification tasks. To deal with the specific problem of audio source separation, an up-sampling layer, block skip connection and band-dedicated dense blocks are incorporated on top of DenseNet. The proposed approach takes advantage of long contextual information and outperforms state-of-the-art results on SiSEC 2016 competition by a large margin in terms of signal-to-distortion ratio. Moreover, the proposed architecture requires significantly fewer parameters and considerably less training time compared with other methods.

Authors

Yuki Mitsufuji

7 papers

Naoya Takahashi

6 papers

TL;DR

Abstract

Authors

References29 items

Improving music source separation based on deep neural networks through data augmentation and network blending

The 2016 Signal Separation Evaluation Campaign

AENet: Learning Deep Audio Features for Video Analysis

Projection-Based Demixing of Spatial Audio

A fully convolutional deep auditory model for musical chord recognition

Multichannel music separation with deep neural networks

Densely Connected Convolutional Networks

Deep Convolutional Neural Networks and Data Augmentation for Acoustic Event Detection

PROJET — Spatial audio separation using projections

Multichannel blind source separation based on non-negative tensor factorization in wavenumber domain

Deep Residual Learning for Image Recognition

Cauchy nonnegative matrix factorization

Very deep multilingual convolutional neural networks for LVCSR

Training Very Deep Networks

Deep neural network based instrument extraction from music

Scalable audio separation with light Kernel Additive Modelling

Deep NMF for speech separation

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

MedleyDB: A Multitrack Dataset for Annotation-Intensive MIR Research

Kernel Additive Models for Source Separation

Exploring convolutional neural network structures and optimization techniques for speech recognition

Deep Sparse Rectifier Neural Networks

Under-Determined Reverberant Audio Source Separation Using a Full-Rank Spatial Covariance Model

Performance measurement in blind audio source separation

“rmsprop adaptive learning,”

Reading Digits in Natural Images with Unsupervised Feature Learning

Learning Multiple Layers of Features from Tiny Images

Département Traitement du Signal et des Images

Gradient-based learning applied to document recognition

Field of Study

Journal Information

Name

Page

Venue Information

Name

Type

URL

Alternate Names