Comparison of filtered back projection with fast non-local means denoising approach and iterative reconstruction in pediatric chest CT image using 3D printed lung nodules

Abstract

This study aimed to quantitatively analyze the efficiency of the fast non-local means (FNLM) filter in increasing the nodule detection sensitivity in pediatric chest computed tomography (CT) using 3D-printed lung nodules. For that purpose, we compared filtered back projection (FBP) with FNLM filter with iterative reconstruction (IR) method. The performance of the proposed FNLM filter was compared through various quantitative evaluations by modeling the previously used noise reduction methods. When the FNLM filter was applied to the acquired FBP reconstruction method-based CT image, the coefficient of variation and contrast-to-noise ratio values in the lung nodule region showed similar values to those of the IR method. In addition, it was demonstrated that the point spread function value that can evaluate sharpness can be improved by using the FNLM filter. In conclusion, the results of this study are expected to maximize the image quality and reduce the dose by fusing the CT image reconstructed by the FBP method and the FNLM filter with excellent characteristics.