Skull Stripping and Classification Of Autism Spectrum Disorder Using Deep Learning

The most popular imaging technique for examining the human brain is neuroimaging. It can assist in the diagnosis and prevention of various illnesses as well as help us comprehend the intricate neuronal characteristics of the human brain. In recent years, deep neural networks have become more and more important in the fields of pattern recognition and machine learning. The best candidate algorithm for picture recognition is convolutional neural networks since they can learn intricate and non-linear mappings from large datasets.ASD (autism spectrum disorder) is a developmental impairment caused by brain differences. The goal of this study is to classify ASD from Normal controls of MRI images using CNN. The dataset used was collected from an organization, such as Autism Brain imaging dataset (ABIDE 1). The dataset collected was a raw image and was preprocessed by applying preprocessing techniques like skull stripping, slicing, Histogram equalization, normalization, and rescaling. This preprocessed data was sent to CNN. By using CNN model, it was able to classify the ASD from Normal controls. The model gave an average accuracy of 95.39%