Research On New CT Data Sampling Techniques With Image Reconstruction Algorithms

CT has been widely used in medical diagnosis,radiotherapy and industrial detection,by virtue of its superiority in accurately providing the faulted images of the scanned object.However,CT systems still have limitations to be solved,such as limited scanning field and dose accumulation of conventional CT system,complex technical requirements and high system cost of multi-energy CT system.Therefore,how to effectively solve the above limitations have important practical values.The theoretical basis of CT imaging,classical data sampling techniques and reconstruction algorithms are firstly reviewed in this paper.In view of the unsuccessful reconstruction with limited scanning field through conventional CT,and the low cost realization of multi-energy CT system.Two new CT data sampling techniques with corresponding reconstruction algorithms are proposed and realized respectively:Firstly,a block fan-beam scanning based parse sampling technique is proposed,to overcome the limitations of scanning field of view in conventional CT system and the harm of excessive cumulative dose to human health.In the technique of block fan-beam scanning based sparse angle sampling,a detector block arrangement is defined and sparse angle projection collection is used to reduce the cumulative dose while collecting effective projections.Meanwhile,an iterative reconstruction method based on proposed data sampling technique is designed to ensure the image quality.According to the simulated and real data experiments',results,this method could expand the scanning field and reconstruct high-quality CT images at a lower dose.The SSIM is up to 0.99,the CNR and MSE increases 6 times and decreases 95%respectively,When the size of detector is limited and the total size of detector remains unchanged.Secondly,a multi-energy sparse segmental data sampling technique and its corresponding iterative reconstruction method is presented to solve the problem of the high requirement of voltage switching frequency of X-ray bulb for fast tube voltage switching based energy spectrum CT system.In this method,a sparse segmental scanning mode is proposed to greatly reduce the voltage switching frequency of the X-ray bulb,which can be realized on the conventional CT system.At the same time,a prior image based joint constraint iterative reconstruction method is designed to effectively retain structural information such as edges while suppressing noise and artifact.Simulated and real data experiments validated this method,and results show that this method can effectively reduce the voltage switching frequency and accurately reconstruct the energy spectrum CT images.Compared with traditional methods,the SSIM is up to 0.98 on average,the CNR increases about 5 times,and the MSE decreases about 56%on average.In this paper,two novel data sampling techniques with reconstruction algorithms are proposed.The former can reconstruct high-quality tomography image with low dose level when the measured object is beyond the range of field of view,and the latter decreased technical requirements of the the tube voltage switching based multi-energy CT.Preliminary achievements in method implementation have been obtained in this paper,but still need further explore for clinical practices.