- Version
- Download 77
- File Size 884.03 KB
- File Count 1
- Create Date 26/07/2022
- Last Updated 26/07/2022
Let me predict your emotions
Mr. Safeer ulla1, Prof. Amos R2
1Mr. Safeer ulla Department of MCA, Maharaja Instituted of Technology Mysore.
2Prof. Amos R, Assistant Professor, Department of MCA, Maharaja Instituted of Technology Mysore.
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract -This As the working of the project the process of speech recognition based on the human emotions or the speech prediction based on the different kinds of the human emotions. It takes the audio file where it contains single line of sentence that in the form of audio signals and then it translated into text by using python library called speech recognition and after converting into the text, we can get to know what is that audio file and apply the masking over it. masking is a technique that clean the background sound in the audio.As a sample of input where in a form of audio which is recorded by the expert or a trainer for the prediction of the model and accuracy level of the information in which form of emotion that audio has been recorded based on the machine and debt learning algorithms the main purpose of developing this kind of project is to predict the emotion of human under the critical situations such as automatic car emotion prediction.Fields such as integrative speech based-agents or caller interaction analysis. The best example of this kind of system used to talk or caller- agent conversation analysis where caller-agent never communicated in same manner where the way of predicting the pitching/commination way of talking is to be consider by one customer to another customer. How is to predict it so we have implemented the system to solve this kind of problem’s where the SEP (speech emotion prediction) system, based on the different type of feature extraction models have been used and developed such as MFCC (Mel-frequency cepstrum coefficients, chroma and Mel and the MLPClassification classification model which we developed in our project to map the suitable predicted emotions and gives the predicted emotion as a result. Where there are different type of feature extraction method and classification models also valuable in the python but using MLPClassification is based on the classification and more suitable so the MLPClassification is the best one and feature extraction models as well with the high accuracy. By using python library called speech recognition and after converting into the text, we can get to know what is that audio file and apply the masking over it. masking is a technique that clean the background sound in the audio. Speech emotion prediction is a system that take a set of audios as the input and predict the emotion based of the sound and pitch and tells the emotion based on the audio recorded for a particular audio have predicted particular results by using different methodologies from machine learning and deep learning algorithms
Key Words:Speech Emotion Prediction, Machine learning, Chroma, Mel, Deep Learning, MLPClassification.