Multi-Speaker Localization Using Convolutional Neural Network Trained with Noise

Abstract

The problem of multi-speaker localization is formulated as a multi-class multi-label classification problem, which is solved using a convolutional neural network (CNN) based source localization method. Utilizing the common assumption of disjoint speaker activities, we propose a novel method to train the CNN using synthesized noise signals. The proposed localization method is evaluated for two speakers and compared to a well-known steered response power method.

Presented at Machine Learning for Audio Processing (ML4Audio) Workshop at NIPS 2017

References in corpus (2)

• Adam: A Method for Stochastic Optimization
• Broadband DOA estimation using Convolutional neural networks trained with noise signals

Cited by in corpus (9)

• Deep Learning for Audio Signal Processing
• Sound Event Localization and Detection of Overlapping Sources Using Convolutional Recurrent Neural Networks
• A Survey of Sound Source Localization with Deep Learning Methods
• Multi-Speaker DOA Estimation Using Deep Convolutional Networks Trained with Noise Signals
• Polyphonic Sound Event Detection and Localization using a Two-Stage Strategy
• Signal-Aware Direction-of-Arrival Estimation Using Attention Mechanisms
• A Deep Reinforcement Learning Approach to Audio-Based Navigation in a Multi-Speaker Environment
• Multi-scale aggregation of phase information for reducing computational cost of CNN based DOA estimation
• A Deep Reinforcement Learning Approach for Audio-based Navigation and Audio Source Localization in Multi-speaker Environments