Random Deep Belief Networks for Recognizing Emotions from Speech Signals.

Guihua Wen,Huihui Li,Jubing Huang,Danyang Li,Eryang Xun

doi:10.1155/2017/1945630

Abstract

Now the human emotions can be recognized from speech signals using machine learning methods; however, they are challenged by the lower recognition accuracies in real applications due to lack of the rich representation ability. Deep belief networks (DBN) can automatically discover the multiple levels of representations in speech signals. To make full of its advantages, this paper presents an ensemble of random deep belief networks (RDBN) method for speech emotion recognition. It firstly extracts the low level features of the input speech signal and then applies them to construct lots of random subspaces. Each random subspace is then provided for DBN to yield the higher level features as the input of the classifier to output an emotion label. All outputted emotion labels are then fused through the majority voting to decide the final emotion label for the input speech signal. The conducted experimental results on benchmark speech emotion databases show that RDBN has better accuracy than the compared methods for speech emotion recognition.

Highlights

Emotions accompany human being in the life everywhere and every moment [1]
This paper presents a novel random deep belief network (RDBN) method for speech emotion recognition, which is composed of the random subspace, DBN, and support vector machine (SVM) within the framework of ensemble learning
This paper presents a random deep belief network (RDBN) ensemble method for speech emotion recognition

Summary

Introduction

Emotions accompany human being in the life everywhere and every moment [1]. They can be recognized and communicated through speech signals that constitute 38% of the whole communicated emotions [2]. Speech emotions tend to have overlapping features, making it difficult to find the correct classification boundaries To deal with these issues, deep learning methods can be considered that can automatically discover the multiple levels of representations in speech signals. Typical classification methods used for speech emotion recognition include hidden Markov model (HMM) [14], Gaussian Mixture Model (GMM) [15], artificial neural networks such as recurrent neural network (RNN) [16], support vector machine (SVM) [17, 18], and the fuzzy cognitive map network [19]. This paper presents a novel random deep belief network (RDBN) method for speech emotion recognition, which is composed of the random subspace, DBN, and SVM within the framework of ensemble learning.

Related Work

Deep Belief Networks

Random Deep Belief Networks for Ensemble

Experiments and Validation

Findings

Conclusion and Future Work