Performance Comparison of Carotid Artery Intima Media Thickness Classification by Deep Learning Methods

Deep learning is a machine learning sub-field that uses deep neural networks. Instead of customized algorithms for each study in this field, it is aimed to cover the wider data set of solutions based on learning the data. Deep learning is a promising approach to solving artificial intelligence problems in machine learning. Nowadays, deep learning algorithms have begun to show themselves in many applications also being studied in biomedical fields. In this medical image processing study, Carotid Artery Intima Media Thickness Ultrasound images were used. Carotid Artery is a type of cardiovascular disease that can result in stroke. If stroke is not diagnosed early, it is in the first place among the disabling diseases. On the other hand, it is the third most common cause of death after cancer and heart disease. For an early diagnose, biomedical image classification performances of VGGNet architecture, which had successful results in the Imagenet competition and an original convolutional neural network model were compared in this study. 501 ultrasound images from 153 patients were used to test the models’ classification performances. It is seen that VGG16, VGG19 and CNNcc models achieved rates of 93%, 90% and 89.1% respectively. These results showed that deep architectures can provide proper classification on biomedical images and this can help clinics to diagnose the disease.