COVID-19 Detection on CT Scans Using Local Binary Pattern and Deep Learning

Serte, S.; Al-Turjman, F.

X-ray and CT scans show lungs, and images can be used to differentiate positive and negative cases. Analyzing these scans using an artificial intelligent method might provide fast and accurate COVID-19 detection. In this paper, a local binary pattern based deep learning method is proposed for the detection of COVID-19 infection on CT Scans. The proposed technique generates local binary pattern (LBP) representations of the CT scans, and then these representations are modeled using fine-tuned models. The fine-tuned models are AlexNet, VGG, ResNet-18, ResNet-50, MobileNetV2, and DensNet-121. We show that the proposed local binary pattern based deep learning model provides higher performance than classic deep learning models for COVID-19 detection. The classification performance of the method provides 90 % AUC value for COVID-19 detection. © 2021, ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering.