Evaluation of reconstructed images from sparse data on the micro-CT system

Abstract

The quantitative estimation was performed to verify the effect the low dose imaging with fewer projection data using micro-CT with total variation (TV) minimization method. To assess the image contrast and noise, sparse data were obtained from the projection data of the phantom and the mouse. Filtered-backprojection (FBP) images of the projection data of the phantom and the mouse were used as a reference image reconstructed with 400 projection data. Universal quality index (UQI) was used as an image similarity metric for comparison between TV minimization algorithm and FBP algorithm for both the phantom and the mouse image. Contrast-to-noise ratio (CNR) values for the image derived from TV minimization algorithm with 80-view was approximately 45 % higher than FBP with 400-view up to 10 % iodine solution material. UQI values for the phantom and the mouse images were measured with 0.974 and 0.999, respectively. Low dose image using TV minimization algorithm was proved to be advantageous for micro-CT system. Contrast and noise properties on image from TV minimization at few-view were higher than FBP 80-view and full-view scan. However, CNR from 15 % iodine solution to 20 % iodine solution were lower than FBP due to the lower-intensity of the x-rays. The reconstructed images for both phantom and mouse from TV minimization algorithm were well matched to FBP-reference images.