RESEARCH

Acoustic & Speech

Investigates various Speech signal processing schemes for acoustic modeling so that more robust speech recognition can be achieved. Our aim is to perform the state-of-art research providing effective means for achieving:

Source Separation Method for Overlapped Acoustic Event Detection

Admin    2021-04-08    👁️ 900

Contents

1.      Introduction

2.      Relevant algorithms

3.      Application demo

1. Introduction
- This research aims at processing an overlapped sound for acoustic event detection (AED) system by means of the source separation technique. The proposed system can be applied in various applications. (e.g. Home/Office surveillance system, Personal life-logging)

2. Relevant algorithms

NMF non-negative Matrix Factorization (NMF)

• NMF algorithm decomposes a matrix V into two nonnegative matrices

• In the acoustic separation problem using NMF, a nonnegative matrix V can be regarded as a magnitude spectrogram

• Basis W and time evolution H are generated by decomposing V

• Update algorithm of the NMF

3.  Application Demo