click here to download the abstract project of emotion recognition
click here to download the base paper of emotion recognition
ABSTRACT
We provide emotion recognition project in this paper. Communication through voice is one of the main components of affective computing in Computing in human-computer interaction. In this type of interaction, properly comprehending the meanings of the words or the linguistic category and recognizing the emotion included in the speech is essential for enhancing the performance. In order to model the emotional state, the speech waves are utilized, which bear signals standing for emotion such as happy, sad, fear, neutral. This project is aiming to design and develop speech based emotion reaction (SER) prediction system, where different emotions are recognized by means of Convolutional Neural Network (CNN) classifiers. Spectral features extracted is mel-frequency cepstral (MFCC). Librosa package in python language is used to develop proposed algorithm and its performance is tested on taking Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS) samples to differentiate emotions such as happiness, surprise, anger, neutral state, sadness, fear etc. Feature selection (FS) was applied in order to seek for the most relevant feature subset. Results show that the maximum gain in performance is achieved by using CNN.
INTRODUCTION
The human brain is an intricate organ that has been a lasting inspiration for research in Artificial Intelligence (AI). The neural networks in brain had the capability of learning all concepts from experiencing low level information and is remembers them which are processed by sensory periphery. Learning language,understanding speech, and recognizing faces are some examples that manifest the remarkable power of the human brain in learning high-level concepts. The main goal of AI is to develop intelligent systems that are able to generate rational thoughts and behaviour similar to human thoughts and performance. Emotion plays a significant role in daily interpersonal human interactions. There are several modalities for expressing human emotions like body-posture, facial expression & voice. Out of which speech is very significant in expressing emotions. In order to communicate effectively with people, the systems need to understand the emotions in speech. A lot of machine learning algorithms have been developed and tested in order to classify these emotions carried by speech. The aim to develop machines to interpret paralinguistic data like emotion, helps in human-machine interaction. The approach for speech emotion recognition (SER) primarily comprises two phases known as feature extraction and features classification phase. The first phase Feature extraction is the key part in the Speech Emotion Recognition. The quality of the features directly influences the accuracy of classification results. Typically, the Feature Extraction method designs handcraftfeatures based on acoustic features of speech. The second phase includes feature classification using linear and non-linear classifiers. The most commonly used linear classifiers for emotion recognition include the Maximum Likelihood Principle (MLP) and Support Vector Machine (SVM) and Convolution Neural Network (CNN). Usually, the speech signal is considered to be non-stationary. Hence, it is considered that non-linear classifierswork effectively for SER.
