Publication Date: 2023/11/16
Abstract: An extensive range of social and communication challenges, frequently accompanied by repetitive behaviors, describe the complicated neurodevelopmental disease known as autism spectrum disorder (ASD). Improving the quality of life for people with ASD requires early detection and intervention. The use of computer vision and machine learning techniques to aid in the early diagnosis and evaluation of autism has attracted increasing interest in recent years. Convolutional neural networks (CNNs) and facial landmarks are used by the FERS to extract pertinent face features after first recording facial expressions using picture or video input. The algorithm then uses a machine learning classifier to forecast autism severity using the extracted emotional cues. Accuracy, precision, recall, and F1-score are a few of the measures used to gauge the classifier's performance in terms of identifying people with ASD. The suggested FERS has the potential to provide a number of advantages, including early autism detection, objective evaluation of social and emotional behaviors, and aiding medical practitioners in making educated judgments about diagnosis and intervention tactics. Additionally, it might offer a useful tool for tracking the development of people with ASD through time.
Keywords: Autism Spectrum Disorder,Convolutional Neural Networks.
DOI: https://doi.org/10.5281/zenodo.10142920
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT23NOV466.pdf
REFERENCES