Image Quality Assessment on Microwave Tomography Based on Filtered-Back-Projection

Abstract

In this paper, the quality assessment of image reconstructed based on filtered-back-projection (FBP) on microwave tomography is presented. It is a common to use the algorithm of back-projection for image reconstruction in classical tomography where the imaging beam passes the object in a straight line. Here, the quality assessment is conducted for the image of cross-sectional wood of tree trunk reconstructed from the measured data at some frequency bands of microwave region, i.e. L-band, S-band, and C-band. Two different techniques of quality assessment, namely Mean Squared Error (MSE) and Structured Similarity Indexing Method (SSIM), are proposed to assess the image quality. The assessment results show that the use of FBP could produce better quality for both MSE and SSIM techniques. While the use of different frequency bands is insensitive to the result of quality assessment.