A Hybrid Method of Image Restoration and Denoise of CT Images

Image degradation is a universal problem in imaging system for the hardware restriction. Digital image processing can be used to enhance resolution and noise reduction effectively; thus to improve image quality with less cost. Aiming at reducing noise in CT imaging, this paper proposes a hybrid algorithm with several improvements on existing algorithms including the constrained least-square-filter algorithm, the Lucy-Richardson algorithm, non-local-means-filter algorithm, and wavelet-filter algorithm. Comparing the results with these algorithms, we conclude that the algorithm combining with Lucy-Richardson algorithm and non-local means filter algorithm achieves better performs in de-noising and thus raising image resolution. This proposed hybrid algorithm for super-resolution images can apparently increase image resolution from 5lp/cm to 11.5lp/cm and significantly reduce mean standard error from 92 to 18.