CT image denoising based on sparse representation using global dictionary

Low-dose CT (LDCT) images tend to be severely degraded by mottle and streak-like noise, and how to enhance image quality under low-dose CT scanning has attracted more and more attention. This work aims to improve LDCT abdomen image quality through a dictionary learning based de-noising method and accelerate the training time at the same time. The proposed method suppresses noise through reconstructing the image use only one dictionary. Experimental results show that the proposed method is effective in suppressing noise while maintaining the diagnostic image details with much more less time.