| Summary: | The COVID-19 pandemic has attracted the attention of big data analysts and artificial intelligence engineers. The classification of computed tomography (CT) chest images into normal or infected requires intensive data collection and an innovative architecture of AI modules. In this article, we propose a platform that covers several levels of analysis and classification of normal and abnormal aspects of COVID-19 by examining CT chest scan images. Specifically, the platform first augments the dataset to be used in the training phase based on a reliable collection of images, segmenting/detecting the suspicious regions in the images, and analyzing these regions in order to output the right classification. Furthermore, we combine AI algorithms, after choosing the best fit module for our study. Finally, we show the effectiveness of this architecture when compared to other techniques in the literature. The obtained results show that the accuracy of the proposed architecture is 95%.
|
|---|
