Abstract:Automatically detecting emotional state in human speech, which plays an effective role in areas of human machine interactions, has been a difficult task for machine learning algorithms. Previous work for emotion recognition have mostly focused on the extraction of carefully hand-crafted and tailored features. Recently, spectrogram representations of emotion speech have achieved competitive performance for automatic speech emotion recognition. In this work we propose a method to tackle the problem of deep features, herein denoted as deep spectrum features, extraction from the spectrogram by leveraging Attention-based Bidirectional Long Short-Term Memory Recurrent Neural Networks with fully convolutional networks. The learned deep spectrum features are then fed into a deep neural network (DNN) to predict the final emotion. The proposed model is then evaluated on the Interactive Emotional Dyadic Motion Capture (IEMOCAP) dataset to validate its effectiveness. Promising results indicate that our deep spectrum representations extracted from the proposed model perform the best, 65.2% for weighted accuracy and 68.0% for unweighted accuracy when compared to other existing methods. We then compare the performance of our deep spectrum features with two standard acoustic feature representations for speech-based emotion recognition. When combined with a support vector classifier, the performance of the deep feature representations extracted are comparable with the conventional features. Moreover, we also investigate the impact of different frequency resolutions of the input spectrogram on the performance of the system.

Feature Analysis on Emotional Chinese Whispered Speech

A Preliminary Study on Emotions of Chinese Whispered Speech

Tone Recognition of Chinese Whispered Speech

Deep Spectrum Feature Representations for Speech Emotion Recognition

Research on Tone Features of Chinese Vowel Whispers

Acoustic Analysis of Emotional Speech in Mandarin Chinese

Tone Features in Whispered Chinese

Acoustic Analyses of Whispered Mandarin

A Preliminary Study On Vocal Tract System Of Chinese Whispered Vowels

Development of Chinese whispered database for speaker verification

Acoustic Analysis of Whispery Voice Disguise in Mandarin Chinese

Performance Analysis Of Mandarin Whispered Speech Recognition Based On Normal Speech Training Model

Speaker factor analysis of whispered speech from global spectral features

Speaker Identification of Whispered Speech with Perceptible Mood

Feature Selection Filtering Methods for Emotion Recognition in Chinese Speech Signal

Perceptual Study for Emotional Speech of Mandarin Chinese

A Study on Emotional Feature Recognition in Speech.

Contribution of Glottal Waveform in Speech Emotion: A Comparative Pairwise Investigation

Study on Emotion Feature Analysis and Recognition in Speech Signal: an Overview

A Study of Correlation Between Physiological Process of Articulation and Emotions on Mandarin Chinese.

EEMD-Based Speaker Automatic Emotional Recognition in Chinese Mandarin