Method and device for noise reduction of image

Abstract

The invention discloses a method and device for noise reduction of an image. The method for noise reduction of the image comprises the steps that the image I0 is input, noise reduction based on weighted average is conducted on the image I0, an initial smooth image I1 and a weighted image W1 are obtained, and the weighted image W1 is composed of smooth weights W1 (x, y) of all pixels in the smooth image I1; a target weight Wt is selected based on the weighted image W1; the smooth weight W1 (x, y) of each pixel and the target weight Wt are compared, the pixels with the smooth weights W1 larger than the target weight Wt are corrected reversely, and the pixels with the smooth weights W1 smaller than the target weight Wt are corrected forwardly. According to the method and device for noise reduction of the image, in combination with the smooth weights of all the pixels of the image and the target weight obtained based on the smooth weights, a region which is excessively smooth and a region where noise reduction is incomplete in the image are selected, and the two regions are processed respectively; as a result, noise is reduced effectively, the problem that the smoothness degrees of parts of the image are not uniform in the noise reduction process of the image is solved, and noise reduction of the image is more uniform.

Description

technical field [0001] The invention relates to the field of image processing, in particular to an image noise reduction method and device. Background technique [0002] Image noise has attracted more and more attention in digital image processing technology. For example, in the medical field, image noise is also one of the important factors determining the quality of medical images, which is directly related to the effect of diagnosis and treatment. Medical images are inevitably polluted by random noise (such as human body thermal noise, mechanical noise, etc.) during the generation process, so noise reduction has become an important step to improve image quality. [0003] A major difficulty in noise reduction is controlling the degree of smoothing. Too low a level of smoothness will result in insufficient noise reduction, and too high a level of smoothness will result in a loss of image detail, which is visually unnatural. Denoising filters in traditional image denoising...

AI Technical Summary

Problems solved by technology

[0007] The invention aims to solve the technical problem of uneven noise reduction in medical im

Images