Detection Attention Deficit Hyperactivity Disorder by using Convolution Neural Network

Document Type : Original Article

Authors

1 ELECTRONIC AND COMMUNICATION DEPARTMENT, FACULTY OF ENGINEERING, MENOUFIA UNIVERSITY, EGYPT.

2 Communications Department, Faculty of Electronic Engineering, Menoufia University, Faculty of Engineering, 6-October University

Abstract

Attention deficit hyperactivity disorder (ADHD) is a neurological disease that is very common in recent times, and many attempts have been made to overcome it. ADHD is diagnosed in boys more than girls. Girls are more likely to have only symptoms of inattention, and less likely to exhibit disruptive behavior that makes ADHD symptoms more noticeable. This means that girls with ADHD may not always be diagnosed. Artificial intelligence has played a very important role in eliminating this disorder using deep learning technology.  Deep learning has three algorithms as Deep Neural Network (DNN), convolution neural network (CNN), Recurrent Neural Network (RNN). The disease is diagnosed using functional magnetic resonance imaging (fMRI) to determine whether the person is affected or not by taking some snapshots of brain images. A convolutional neural network (CNN) was chosen to extract the specifications or features of fMRI images.There were an optimization technique of the fMRI datasets namely, Nesterov-Accelerated Adaptive Moment Estimation (Nadam). Using these optimization techniques for adapting the classification system for three CNN network or models for ADHD cases, it was concluded that the accuracy for CNN NET 1 is 97.5%, accuracy for CNN NET 2 is 95% and accuracy for CNN NET 3 is 98.75 %. Finally, it’s found that CNN NET 3 is the best as its high accuracy so the system is improved

Keywords

Files

Share

How to cite

Statistics