Application of Sinogram-Based Moving Window Subtraction Technique in OSEM Algorithm to Reduce Ring Artifacts for Cone-Beam Micro-CT Image Quality Assessment

摘要

Severe ring artifacts in cone-beam micro-CT degrade image quality and usually cause error recognition during image-based-observer study or computed-aid-diagnosis (CAD). It causes wrong values, such as size of the tumor. The aim of this study was to derive a moving window subtraction method before the iterative reconstruction step to reduce ring artifacts. To access acceptable image quality data, we used a uniform water cylinder phantom and followed the performance measurement guide, ISO-15708 and EN 16016, to evaluate the imaging performance. Results show there are 1.23 times improvement in SNR and 5% improvement in uniformity. Most importantly, we can reduce the error estimate of the object size from 6 pixels to less than 2 pixels. Our proposed method can efficiently reduce the error rate caused by ring artifacts for automatic size estimation applications.