Improved total-variation noise-reduction technique with gradient method using iteration counter and its application in medical diagnostic chest and abdominal X-ray imaging

Abstract

The development of novel noise-reduction algorithms is one of the main research areas in the field of medical diagnostic X-ray imaging. The purpose of this study is to develop an improved total-variation (TV) noise-reduction technique using a gradient method that employs an iteration counter, and confirm its feasibility for applications in chest and abdominal X-ray imaging systems. We evaluated image qualities using contrast-to-noise ratio (CNR) and relative standard deviation (RSD) calculations with the proposed and conventional noise-reduction algorithms for both chest and abdominal X-ray images. The results of the quantitative CNR and RSD evaluations demonstrated the superiority of the proposed TV noise-reduction technique over other algorithms; in particular, the comparison between the conventional and proposed TV techniques revealed that the proposed method has improved noise-reduction characteristics in all cases. In conclusion, the results demonstrated that the improved TV noise-reduction technique with the gradient method using the iteration counter can be used to acquire images with a better quality than those of conventional algorithms. This can lead to lowering of the radiation dose in chest and abdominal X-ray imaging systems.