Analysis of 3D Linear and Non-linear Filtering Effects Based on 3D MDCT Abdominal Images

摘要

The X-ray exposure is high in the imaging of 3D MDCT (Multi-detector Raw Computed Tomography) methods. Although the effective way of decreasing this hazard is to lower down the dose level of X-ray, it degrades the obtained images and causes negative effects on medical diagnosis, especially in the abdominal region. In this paper, the correlations among voxels in the 3D space are used to suppress the noise. 3D linear and non-linear filters are designed according to the lesion characteristics and the 3D voxel correlations. 3D combination filter is applied according to the advantages of linear and non-linear filters. Experimental results show the developed filtering methods can improve the low dose MDCT images without compromising the diagnostic quality of images.