Automatic multi-organ labeling in thoracic low-dose CT images with the internal dosimetry prospective using the pix-to-pix GAN deep learning method (preliminary study)

، صفحه 0-0 (1)
عنوان دوره: اولین دوره بین المللی و بیست و هشتمین دوره ملی (1400)
The Monte Carlo method plays a significant role in internal dosimetry and has led to the development of accurate patient-specific approaches. Regarding the facts that vowelized patient phantom is a crucial part of the simulation, organ segmentation based on CT images must be performed accurately and automatically. This study aimed to survey an automatic multi-organs segmentation method using low-dose CT images based on the pix-to-pix GAN network. Evaluation metrics, such as DSC and MSD, were assessed to estimate the similarity of the output to ground truth images, and close matching was found between organ automatic delineation and manual contouring.
کلیدواژه ها