A DFC taxonomy of Speech emotion recognition based on convolutional neural network from speech signal

摘要

Speech is an efficient agent to explicit attitude and emotions via language. The crucial task for the researchers is to find out the emotions through the speech utterance and eliminating the noise from a raw speech data. The goal of this research paper is to explore the latest journal papers in the field of convolutional neural network-based speech emotion recognition (SER) models related with the specific problem and provide a best solution which can recognize emotion in the speech from the speech signal.The components of this proposed system are data, feature extraction and classification (DFC) that helps to assist in the implementation and evaluating the system. We propose the DFC taxonomy which will assist the end users in recognition of the emotion from the speech signal and making the artificial intelligence (AI) more robust by using convolutional neural network, facilitating a huge presence in the future system.The system evaluates a state-of-the-art model that is associated to the convolutional neural network-based speech emotion recognition which presents and validates the DFC components. Based on system completeness, system acceptance, and by classifying 30 state-of-the-art journal research papers in the domain, components are evaluated, verified and validated.The benefaction of this research paper is the critical analysis in the latest literature that are available on the convolutional neural network-based system which can recognize the emotion by extracting the features from the speech signal so that accurate recognition of emotion can be made. Also, highlighting the importance of DFC taxonomy.