Applying Iterative Reconstrucion Algorithms In Stationary Digital Breast Tomosynthesis

Breast cancer is the most frequently diagnosed cancer in women worldwide.Due to the increasing growth rate,diagnosed in time has been one of important demands today.Early breast cancer checking is using a normal mammography,but because it is a way to reconstruct a 3-demensions body to a 2-demensions projection,the overlapping problem has been a technique problem before.With the development of technique and deeper research,digital breast tomosynthesis(DBT),which is based on limited-angle tomography,solves this problem but also has limitations.The single X-ray source must rely on the mechanical rotation to achieve multi-angle projection that may easier to get motion artifacts in image due to the machine wobbled.We developed and demonstrated a stationary digital breast tomosynthesis(s-DBT)system that uses a carbon nanotube(CNT)X-ray source array for this problem.Since s-DBT is still in the research stage,few algorithms are used today,and it is also an interesting question whether the placement of the light source will affect the projection.The main work of this paper is to apply the iterative algorithm to the new system and evaluate it.At the same time,the algorithm is improved based on the results of actual experiments and simulation experiments.In simulation experiments,to determine the influence from two types of arrays—a linear array and an arc array—to accurately compare the different reconstruction algorithms and get the highest image quality.In this paper,we used reconstruction algorithms,including the Feldkamp-Davis-Kress(FDK)algorithm,statistical iterative reconstruction technique(SIRT),and simultaneous algebraic reconstruction technique(SART),to reconstruct images from the projection images obtained from the s-DBT and the simulated experiments,verified the feasibility of image reconstruction algorithms using in the stationary digital breast tomosythesis.To acquire the best algorithm for reconstructing the results and the best source array,the peak signal-to-noise ratio(PSNR),root-mean-square error(RMSE),and structural similarity(SSIM)were used in the experiment to analyze the data in the reconstruction algorithms.Finally,improve the algorithm using in s-DBT for its limited application.After double verification of simulation experiments and actual experiments,the iterative method is better applied to static mammary tomography.In simulated experiments,different arrays have minimal effect on the image quality,so considering the linear light source array from practical use may be a more convenient choice.Differ from the traditional DBT system,from all of the experiments,in parameter analysis,SIRT performs better in reconstruction regardless of whether the phantom has few or many details.In the last,the feasibility of the improved algorithm is verified,improved algorithm suppresses noise better and the reconstruction image quality of SIRT is also improved.