A low-dose CT image simulation method

Abstract

The invention provides a low-dose CT image simulation method. The low-dose CT image simulation method comprises the steps of acquiring original high-dose CT images and original low-dose CT images which are not in one-to-one correspondence; preprocessing the original CT image to obtain corresponding original CT image data respectively; constructing a CT image simulation network model composed of ageneration network and a discrimination network, and initializing network parameters; inputting the original CT image data into the CT image simulation network model, and alternately training the generation network and the discrimination network until optimal parameters of the network are obtained and stored; And inputting high-dose CT images into the low-dose CT image simulation network model which is set as the optimal parameter to generate simulation low-dose CT images which are in one-to-one correspondence with the high-dose CT images. The low-dose CT images in one-to-one correspondence with the original high-dose CT images are obtained, data can be provided for development of algorithms such as CT reconstruction and correction, and work such as training, comparison, verification and algorithm optimization can be carried out conveniently.

Description

technical field [0001] The invention relates to the technical field of medical imaging equipment, in particular to a low-dose CT image simulation method. Background technique [0002] Computed Tomography (CT) has the characteristics of fast scanning time and clear images, and can be used for the examination of many diseases. It has become one of the most important examination tools in hospitals. The radiation problem in CT scanning has also been increasingly affected increasingly widespread attention. Therefore, the development of new technologies that can reduce the radiation dose to patients in CT scan detection has become an urgent need in the medical field. [0003] At present, a large number of studies have applied deep learning methods such as convolutional neural networks and generative adversarial neural networks to low-dose CT image reconstruction, and achieved good results. These deep learning methods require a one-to-one correspondence between training data (low...

AI Technical Summary

Problems solved by technology

[0005] When the inventor was researching, he found that the traditional method had at least the following problems: 1, the added low-dose image had a single noise or feature, which could not simulate the actual system situation; 2, the simulation time was long (the simulation time of a CPU computer with 20 nodes generally needs A few days to more than ten days), parameter adjustment is difficult

Images