An improved ring artifact removal approach for flat-panel detector based computed tomography images

Ring artifacts often appear in flat-panel detector based CT images due to the malfunction and mis-calibration of detector elements resulting in stripe artifacts in the sinogram data. In this paper, we propose an improved ring artifacts removal (IRAR) approach by combining multi-resolution analysis and weighted moving average filter methods. Specifically, the multi-resolution analysis method can suppress these high-intensity vertical stripes in the sinogram domain and the moving average filter method is effective to deal with the relatively high-intensity vertical stripes. Experimental results on the data acquired from a commercial micro-CT scanner show that the presented approach can achieve significant gains over the existing methods in the ring artifacts removal while preserving spatial resolution.