The segmentation of lung nodules is an essential research area in the field of lung cancer detection, however achieving high accuracy is a major challenge due to the variability and distinctiveness of these nodules. In order to overcome this challenge, a generative adversarial network (GAN) model has been developed in the current study for lung nodule segmentation. Pre-processing methods, including the use of a Gaussian filter have been implemented to eliminate any artifacts that may have been present in the input computed tomography (CT) image. After pre-processing stage, the image has further aroused to lung lobe segmentation, which involved utilizing deep-joint segmentation techniques to accurately identify the relevant regions of the lung. GAN has been used to perform the lung nodule segmentation, with the model trained to segment lung nodules from the lung lobe image. Performance of the model has been evaluated using metrics namely accuracy, Jaccard-similarity, and Dice coefficient to assess its effectiveness.

Lung lobe segmentation, Image pro-cessing, Lung nodule seg-mentation, Generative Adversarial-Network