Improved image noise level estimation based on segmentation and block processing
<p>Liwei Han</p> · Academic Journal of Computing & Information Science · 2022
Accurate estimation of noise parameters in digital images is of great significance to improve the quality of image processing. Principal component analysis is an important means of image denoising, and the traditional method is to estimate the whole image. Due to the complexity of image content, this paper proposes a preprocessing method based on superpixel segmentation to process the largest possible smooth block in the image. Compared with contrast method, the estimated value of the proposed method is closer to the true value.