ACNN Based Speech Emotion Recognition and Noise Suppression Using Modified Cuckoo Search Algorithm

Speech is the best way to communicate between two people. Since this is the most common method of communication, lot of researches had been inventing from past few years. Even people always like to interface with machines utilizing speech.Based on this, automatic speech recognition has gained up a big momentum in recent years.In this paper, an efficient Speech Emotion Recognition (SER) system using Modified Cuckoo Search Optimization (MCS) based optimization technique with Actor Critic Neural Network (ACNN) is introducing. Apart from that the noise suppression and speech enhancement is also done with ACNN and is compared with the result of Artificial Neural Network(ANN) using other optimization algorithm. Nowadays lot of researches has been done in the emotion recognition and the speech enhancement.The other difficult task behind this emotion recognition is none other than choosing the emotion recognition corpora(speech database), identification of different features related to speech and an appropriate choice of a classification model.The various types of classifiers are used to differentiate emotions. Preprocessing, feature extraction and classification are the important stages in this signal model. Local invariant features and salient discriminative features are extracted in the feature extraction.These features are given to the neural network for training and based on that the text correspond to the speech signal is recognized.This system will be a speaker independent, since it use recorded standard database of speech samples which have different speech emotions.