L.T. ALAYUE, B.S. GOSHU#, ENDRIS TAJU
Department of Physics, Dire Dawa University
Lung cancer is one of the most life-threatening diseases. It is a medical problem that needs accurate diagnosis and timely treatment by healthcare professionals. Although CT is preferred over other imaging modalities, visual interpretation of CT scan images may be subject to error and can cause a delay in lung cancer detection. Therefore, image processing techniques are widely used for early-stage detection of lung tumors. This study was conducted to perform pre-processing, segmentation, and feature extraction of lung CT images using image processing techniques. We used the MATLAB programming language to devise a stepwise approach that included image acquisition, pre-processing, segmentation, and features extraction. A total of 14 lung CT scan images in the age group of 55–75 years were downloaded from an open access repository. The analyzed images were grayscale, 8 bits, with a resolution ranging from 151 213 to 721 900, and Digital Imaging and Communications in Medicine (DICOM) format. In the pre-processing stage median filter was used to remove noise from the original image since it preserved the edges of the image, whereas segmentation was done through edge detection and threshold analysis. The results show that solid tumors were detected in three CT images corresponding to patients aged between 71 and 75 years old. Our study indicates that image processing plays a significant role in lung cancer recognition and early-stage treatment. Health professionals need to work closely with medical physicists to improve the accuracy of diagnosis.
Key words: CT image, lung cancer detection, MATLAB, image processing.
Corresponding author’s e-mail: firstname.lastname@example.org