New Line Contrast Figure of Merit for image quality assessment

Abstract

A novel Figure of Merit (FoM) is presented which is based on analysis of data extracted from line profiles within the image. This FoM is referred as Line Contrast (LC). In order to evaluate its performance, the proposed FoM along with Contrast to Noise Ratio (CNR), are applied in a simulated phase contrast image and the resulting figures are compared. Phase contrast image is created using the XRAYImagingSimulator an in-house developed software platform for x-ray imaging simulation. A semi-cylindrical phantom especially designed for phase contrast applications that contains materials mimicing the refractive properties (indexes) of breast tissues was used. Images were produced at 20 keV with source to isocenter distance (SID) 23m and object to detector distance (ODD) 400mm and pixel size of 5μm. Results from the application of both FoMs shows that LC comes into full accordance with visual assessment and that it overpass the traditionally used CNR, especially in the ability of detecting the edge enhancement due to phase contrast, of the different structures within the image. The edge detection features make it optimal for phase contrast investigation where the effect on the edge enhancement is of outmost importance.