Diagnosis of Multiple Sclerosis Disease in Brain MRI Images using Convolutional Neural Networks based on Wavelet Pooling

Multiple Sclerosis (MS) is a disease that destructs the central nervous system cell protection, destroys the sheaths of immune cells, and causes lesions. Examination and diagnosis of lesions by the specialists is usually done manually on the Magnetic Resonance Imaging (MRI) images of the brain. The factors such as the small sizes of lesions, their dispersion in the brain, similarity of lesions to some other diseases, and their overlap can lead to a misdiagnosis. The automatic image detection methods, as auxiliary tools, can increase the diagnosis accuracy. To this end, the traditional image processing methods and deep learning approaches have been used. The deep convolutional neural network is a common method of deep learning to detect lesions in the images. In this network, the convolution layer extracts the specificities, and the pooling layer decreases the specificity map size. In the present research work, we used the wavelet-transform-based pooling. In addition to decomposing the input image and reducing its size, the wavelet transform highlights the sharp changes in the image and better describes the local specificities. Therefore, using this transform can improve the diagnosis. The proposed method is based on six convolutional layers, two layers of wavelet pooling, and a completely connected layer that has a better amount of accuracy than the studied methods. The accuracy of 98.92%, precision of 99.20%, and specificity of 98.33% are obtained by testing the image data of 38 patients and 20 healthy individuals.