Publication

Abstract : In recent years, Deep Learning (DL) has exhibited prevalent accomplishments in numerous data-rich applications, including healthcare. This work presents a high accuracy deep learning model to detect Schizophrenia or bipolar disorder or neither of them i.e., healthy. This work fosters a productive algorithm to perceive whether an individual is suffering from Schizophrenia disorder or bipolar disorder or is healthy. First, a sequential model was built to understand how accurate the model is in predicting the mental health of aged persons. Using this knowledge, the VGG-16 model is implemented to improve the accuracy of prediction. The limitations of this model are identified and to tackle the limitations VGG-19 model is utilized. The accuracy of the models carried out is obtained as Sequential: 86.05%, VGG-16: 92.10%, and VGG-19: 94.78%, and the developed proved to give better results with 94.78% accuracy using the VGG-19 model.