Mind Care Solution Through Human Facial Expression

Abstract

Using proposed system psychologists can use technology to make decisions which can provide ease for both patients and psychologists. Psychologists can check the progress of patients by analysing emotions reports of patient over time. Using historical data and emotion detection technology psychologists can make more accurate decisions. Using proposed system patient and psychologists don’t have to go to anywhere they only need a device and internet. Based on the characteristics of patient emotion psychologist only need report generated by system and prescribe medicine in emergency situation. Proposed system improves consultancy method by using machine learning emotion detection algorithm. Proposed system detects facial emotion of patient by using CNN with HAAR cascade classifier. We use FER 2013 dataset to train our model. We use VGG 19 architecture to train our model for optimization function to enhance the accuracy of model. We use RELU. We use DJANGO framework for integration with frontend. Result of our model on dataset 82.3% after find tuning the accuracy goes to 82.3% to 92%. We use recall and F1 method to check the performance of model. We trained model on the testing dataset which have gray scale images and 48*48pixel images to achieve his performance. To achieve our accuracy goal, we split dataset into trainee validation and testing dataset. We use CNN and achieve 93% accuracy in our system which help patient to get feedback only selected question and psychologist. Patients select psychologist to answer questions of psychologist system stores emotions of patient against every question to generate emotion report. Psychologist can analyze emotion report to provide better prescription to patient.

Keywords

• emotion detection
• cnn
• feature extraction
• psychologist
• patient