Real Time Facial Emotion Recognition model Based on kernel Autoencoder and Convolutional Neural Network for Autism Childrens

Abstract

Abstract Autism spectrum disorder (ASD) is a developmental disability brought on by abnormalities in the brain. Patients with ASD usually struggle with social contact and communication. They may also have a problem with the traditional ways of learning and paying attention. Diagnosis of autism considers a challenging task for medical experts since the medical diagnosis mainly depends on the abnormalities in the brain functions that may not appear in the early stages of early onset of autism disorder. Facial expression can be an alternative and efficient solution for the early diagnosis of Autism. This is due to Autistic children usually having distinctive patterns which facilitate distinguishing them from normal children Assistive technology has proven to be one of the most important innovations in helping autistic improve their quality of life. A real-time emotion identification system for autistic youngsters was developed in this study to detect their emotions to help them in case of pain or anger. Face identification, facial feature extraction, and feature categorization are the three stages of emotion recognition. A total of six facial emotions are detected by the propound system: anger, fear, joy, natural, sadness, and surprise. This research presents a deep convolutional neural network (DCNN) architecture for facial expression recognition to help medical experts as well as families in detecting the emotions of autistic children. To enhance the algorithm performance to classify the input image efficiently, the proposed algorithm contains an autoencoder for feature extraction and feature selection. Due to the size of the used dataset, a pre-trained model( ResNet, MobileNet, and Xception) is used. The xception model achieved the highest performance (ACC = 0.9523%, sn = 0.932, R = 0.9421, and AUC = 0.9134%). The proposed emotion detection framework takes the benefit of using fog and IoT to reduce the latency for real-time detection with fast response and to be a location awareness. As well as using fog is useful while dealing with big data.