A locally adaptive edge preserving filter for denoising of low dose CT using multi-level fuzzy reasoning concept

To reduce the radiation exposure, low dose CT (LDCT) imaging has been particularly used in modem medical practice. The fundamental difficulty for LDCT lies in its heavy noise pollution in the projection data which leads to the deterioration of the image quality and diagnostic accuracy. In this study, a novel two-stage locally adaptive edge preserving filter based on multi-level fuzzy reasoning (LAEPMLFR) concept is proposed as an image space denoising method for LDCT images. The first stage of structured pixel region employs multi-level fuzzy reasoning to handle uncertainty present in the local information introduced by noise. The second stage employs a Gaussian filter to smooth both structured and non-structured pixel region in order to retain the low frequency information of the noisy image. Comparing with traditional denoising methods, the proposed method demonstrated noticeable improvement on noise reduction while maintaining the image contrast and edge details of LDCT images.