JIITA, vol.6 no.1, p.505-512, 2022, DOI:
Design of Medical Image Information Classifier to Improve the Accuracy of Lung Cancer Diagnosis
Minuk Jeong 1), Yoosoo Oh 1,*)
1) Dept. of ICT Convergence, Dept. of AI, Daegu University, Gyeongsan-si, Republic of Korea
Abstract: The incidence of lung cancer is increasing every year, and the first cause of death due to cancer is lung cancer. From 2012 to 2016, the most misdiagnosed of cancer among medical damage relief applications, and most of the damage cases were cancer but were misdiagnosed as non-cancer. In this paper, we propose a medical image information classifier to improve lung cancer diagnosis accuracy. The proposed classifier serves to assist the user in diagnosing lung cancer by reading medical images. They are data sets obtained from The Cancer Imaging Archive (TCIA) of the National Cancer Institute (NCI), USA. Pre-processing is performed using Houns Field Unit Changes. Then, classifiers are implemented as training using 3D CNN algo-rithms, one of the types of deep learning algorithms. As a result of the im-plementation, the performance of the classifier was 87.8%. Finally, the learned classifier model is applied to the AIoT device, and the medical im-age judgment result is provided to the user through the GUI on the screen.
Keywords: Lung Cancer, Deep Learning, CNN, IoT