Emotion recognition in speech signals using optimization based multi-SVNN classifier

Abstract

Emotion recognition is an interdisciplinary area, and it achieves a significant attention of the researchers in the past few years. Automatic recognition of an emotional state intends to attain an interface among the machines and human beings. Accordingly, a speaker emotion recognition system, named Fractional Deep Belief Network (FDBN), is designed in the literature using fractional deep belief network that combines fractional theory and deep belief network. This work introduced a novel emotion recognition scheme, Whale-Imperialist Optimization algorithm (Whale-IpCA) based Multiple Support Vector Neural Network (Multi-SVNN) classifier, for identifying the emotions in the speech signal. The newly proposed Whale-IpCA algorithm hybridizes the Whale Optimization Algorithm (WOA) and Imperialist Competitive Algorithm (IpCA), and it trains the Multi-SVNN classifier for identifying the emotions. Also, the spectral feature set is extracted from the input signal and provided to the proposed Whale-IpCA based Multi-SVNN for the recognition purpose. Simulation of the proposed Whale-IpCA based Multi-SVNN is done with the help of the standard emotion databases, such as Berlin and Telugu. From the results, it is evident that the proposed Whale-IpCA based Multi-SVNN classifier has surpassed other existing works with the values of 0.0025, 0, and 0.9987 for the FNR, FPR, and accuracy, respectively.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 49 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.