Autism Spectrum Disorder (ASD) is a severe neuropsychiatric disorder that affects intellectual development, social behavior, and facial features, and the number of cases is still significantly increasing. Due to the variety of symptoms ASD displays, the diagnosis process remains challenging, with numerous misdiagnoses as well as lengthy and expensive diagnoses. Fortunately, if ASD is diagnosed and treated early, then the patient will have a much higher chance of developing normally. For an ASD diagnosis, machine learning algorithms can analyze both social behavior and facial features accurately and efficiently, providing an ASD diagnosis in a drastically shorter amount of time than through current clinical diagnosis processes. Therefore, we propose to develop a hybrid architecture fully utilizing both social behavior and facial feature data to improve the accuracy of diagnosing ASD. We first developed a Linear Support Vector Machine for the social behavior based module, which analyzes Autism Diagnostic Observation Schedule (ADOS) social behavior data. For the facial feature based module, a DenseNet model was utilized to analyze facial feature image data. Finally, we implemented our hybrid model by incorporating different features of the Support Vector Machine and the DenseNet into one model. Our results show that the highest accuracy of 87% for ASD diagnosis has been achieved by our proposed hybrid model. The pros and cons of each module will be discussed in this paper.
Currently, every 1 in 54 children have been diagnosed with Autism Spectrum Disorder (ASD), which is 178% higher than it was in 2000. An early diagnosis and treatment can significantly increase the chances of going off the spectrum and making a full recovery. With a multitude of physical and behavioral tests for neurological and communication skills, diagnosing ASD is very complex, subjective, time-consuming, and expensive. We hypothesize that the use of machine learning analysis on facial features and social behavior can speed up the diagnosis of ASD without compromising real-world performance. We propose to develop a hybrid architecture using both categorical data and image data to automate traditional ASD pre-screening, which makes diagnosis a quicker and easier process. We created and tested a Logistic Regression model and a Linear Support Vector Machine for Module 1, which classifies ADOS categorical data. A Convolutional Neural Network and a DenseNet network are used for module 2, which classifies video data. Finally, we combined the best performing models, a Linear SVM and DenseNet, using three data averaging strategies. We used a standard average, weighted based on number of training data, and weighted based on the number of ASD patients in the training data to average the results, thereby increasing accuracy in clinical applications. The results we obtained support our hypothesis. Our novel architecture is able to effectively automate ASD pre-screening with a maximum weighted accuracy of 84%.