Finite angle CT reconstruction artifact removing method based on generative adversarial network

Abstract

The invention discloses a finite angle CT reconstruction artifact removing method based on a generative adversarial network. The method comprises the following steps: performing downsampling and filtering back projection processing on a full-angle CT image in an angle range of [45.5 degrees, 135.5 degrees] to obtain a limited-angle CT image, splicing the full-angle CT image with the limited-angleCT image, and dividing a plurality of spliced images into a training set and a test set as data samples; based on a WGAN-GP network model which comprises a generator and a discriminator, and constructing deep neural network structures of the generator and the discriminator respectively; sending the training set into a WGAN-GP network model for training, and sending test set data into the trained model for testing; and sending the to-be-processed image into the trained model to output a clear finite angle CT reconstruction image. Compared with the prior art, the method for removing the block artifacts is better in removing effect, and meanwhile details and edge information can be well reserved.

Description

technical field [0001] The invention relates to a method for removing artifacts in limited-angle CT reconstruction based on a generative confrontation network, and belongs to the technical field of medical image processing. Background technique [0002] Computed tomography (CT, Computed Tomography) is playing an increasingly important role in clinical diagnosis. However, in the practical application of CT, some data regions cannot be sampled due to the limitation of physical acquisition. Limited-angle CT can quickly scan patients. Although the X-ray dose is reduced and the harm to the body is reduced, the imaging effect is not as outstanding as that of full-angle projection. [0003] In order to reduce the harm to human health, X-ray dose should be strictly controlled in clinical diagnosis. Currently, major manufacturers including General Electric, Toshiba, and Philips are working on clinically lower doses. In general, reducing the tube current (or voltage) and the number...

AI Technical Summary

Problems solved by technology

When the image is similar to the image to be reconstructed, the PICCS algorithm can reconstruct the image well, but when there are obvious differences between the two images, the pseudo-structure information of the prior image is introduced into the reconstructed image, thereby reducing the image quality

Images