Age & gender based voice pathology classification using deep neural technique

Voice pathology can be successfully identified using computer-aided voice pathology categorization approaches. The goal of this research is to develop a tool using deep learning technique to diagnose voice pathological conditions including age and gender aspects. Here, a dataset of vocal pathology is used to train a Convolutional Neural Network (CNN) which is a pre-trained model to improve classification accuracy. The proposed work exhibits a high level of instantaneous, automatic diagnostic and therapeutic alternatives to achieve classification accuracy. Usually, recordings from people of all ages are used to develop algorithms that can recognize speech abnormalities automatically. However, this approach might not be sufficient due to the acoustic changes in the voice brought on by aging naturally. Additionally, it can be more challenging to identify pathological voice since the quality of older voices changes in ways that are comparable to those linked to vocal problems. Research developing methods that automatically account for speaker age to make it easier to identify speech anomalies is therefore of interest. The current study introduces an age/gender based automated classification utilizing both healthy and pathological voice.