Automatic detection and contouring nodules Based on Lung Computed Tomography Images using U-Net Segmentation Method for Treatment Planning System

، صفحه 0-0 (1)
عنوان دوره: اولین دوره بین المللی و بیست و هشتمین دوره ملی (1400)
Deep learning algorithms have recently been developed that utilize imaging information, as a means to increase treatment planning efficiency and improve radiotherapy plan quality. In this study lung nodule region segmented automatically using U-Net that is the popular segmentation network for biomedical images. 358 CT images from patients were used to train and test U-Net. In this research, lung cancer volume calculation process of target based on CT images was done. Calculation of target volume was done in purpose to treatment planning system in radiotherapy. The calculation of the target volume was done by adding the target area on each slices and then multiply the result with the slice thickness. The calculation of nodule region targets is 611 mm2 for area target volume and 1527.5 mm3 for GTV. In this study, the value of dice obtained 83% that confirms the usefulness of the proposed method in treatment planning systems.
کلیدواژه ها