Partially Shared Deep Neural Network in sound source separation and identification using a UAV-embedded microphone array

This paper addresses sound source separation and identification for noise-contaminated acoustic signals recorded with a microphone array embedded in an Unmanned Aerial Vehicle (UAV), aiming at people's voice detection quickly and widely in a disaster situation. The key approach to achieve this is Deep Neural Network (DNN), but it is well known that training a DNN needs a huge dataset to improve its performance. In a practical application, building such a dataset is not often realistic owing to the cost of manual data annotation. Therefore, we propose a Partially-Shared Deep Neural Network (PS-DNN) which can learn multiple tasks at the same time with a small amount of annotated data. Preliminary results show that the PS-DNN outperforms conventional DNN-based approaches which require fully-annotated data in training in terms of identification accuracy. In addition, it maintains performance even when noise-suppressed signals are used for sound source separation training, and partially annotated data is used for sound source identification training.

Authors

• Takayuki Morito
• Osamu Sugiyama
• Ryosuke Kojima
• Kazuhiro Nakadai

Keywords

• Training
• Source separation
• Neural networks
• Machine learning
• Microphone arrays
• Acoustics
• Neural Network
• Deep Neural Network
• Identification Of Sources
• Sound Source
• Microphone Array
• Sound Source Separation
• Sound Source Identification
• Annotation Data
• Unmanned Aerial Vehicles
• Acoustic Signals
• Manual Annotation
• Disaster Situations
• Speech Detection
• Amount Of Annotated Data
• Deep Learning
• Convolutional Neural Network
• Types Of Information
• Classification Task
• Hidden Layer
• Multi-channel Signals
• Clear Signal
• Huge Amount Of Data
• Data In Dataset
• Model Retraining
• Sound Localization
• Inertial Measurement Unit
• Laser Ranging
• Short-time Fourier Transform
• Input Vector
• robot audition