A CT Image Reconstruction Method Based on Offline Dictionary Sparse Regularization for Scanning with Low Tube Current Intensity

Abstract

The invention, which belongs to the technical field of the medical image processing, especially relates to an offline-dictionary-sparse-regularization-based CT image reconstruction method in a state of low tube current intensity scanning. A plurality of existing clear CT images of different parts are taken and are used as a sample set, an offline dictionary is trained, and offline-dictionary-based sparse representations of the CT images are used as regularization items; and under the circumstance of low tube current intensity projection, image reconstruction is carried out by using a statistic iterative reconstruction algorithm. The method has the following beneficial effects: the quality of the reconstructed image can be enhanced on the condition of low-X-ray tube current projection; and a clear reconstructed image with structural details kept can be obtained when the radiation dosage is reduced to be 10% of that of the traditional FBP algorithm or even is reduced to be lower than 10%.

Description

technical field [0001] The invention belongs to the technical field of low-dose CT image reconstruction, in particular to a CT image reconstruction method based on off-line dictionary sparse regularization and low tube current intensity scanning. Background technique [0002] CT technology has the characteristics of fast, accurate, non-invasive, and pain-free, and can clearly obtain the attenuation information of different tissues of the human body for X-rays on the millimeter scale, thereby providing rich three-dimensional human organ tissue information for clinicians' diagnosis and prevention. As a commonly accepted clinical examination method, CT has become one of the indispensable and main tools in the field of radiological diagnosis. Generally, the current intensity of a clinical diagnosis CT scan X-ray tube generally exceeds 100mA. During the transmission process of X-rays, part of the energy will be transferred to the human body, which will cause physical damage, ind...

AI Technical Summary

Problems solved by technology

[0006] Existing reconstruction algorithms based on dictionary sparse representation need to segment the intermediate reconstruction image as a training sample set for dictionary training in each iteration, which requires a large amount of calculation and takes a long time, and the trained dictionary will be affected by intermediate image artifacts. There are still some differences between the results and the real images

However, the sparse representation methods of other natural images are not very specific to the characteristics of CT images.

In addition, the dictionary learning reconstruction algorithm currently proposed in the field of CT image reconstruction cannot reflect the statistical characteristics of projection noise. image reconstruction

Images