METHOD AND SYSTEM FOR GENERATING COMPOSITE PET-CT IMAGE ON BASIS OF NON-ATTENUATION-CORRECTED PET IMAGE

摘要

Disclosed are a method and system for generating a composite PET-CT image on the basis of a non-attenuation-corrected PET image. The method comprises: constructing a first generative adversarial network and a second generative adversarial network; training the first generative adversarial network, so as to obtain a mapping relationship between a PET image which has not been subjected to attenuation correction and a PET image which has been subjected to attenuation correction; training the second generative adversarial network, so as to obtain a mapping relationship between the PET image which has been subjected to attenuation correction and a CT image; and generating a composite PET/CT image by using the obtained mapping relationships. By means of the present invention, a high quality PET/CT image can be directly compounded from a non-attenuation-corrected PET image, the medical expenses of a patient can be reduced, and the radiation dose received by the patient during an examination process is minimized.