Fractional‐integral‐operator‐based improved SVM for filtering salt‐and‐pepper noise. Issue 12 (18th September 2019)

• Record Type: Journal Article
• Authors: Jia, Xiaofen; Guo, Yongcun; Zhao, Baiting; Huang, Yourui
• Is Part Of: IET image processing. Volume 13:Issue 12(2019)
• Page Start: 2346
• Page End: 2357
• Publication Date: 2019-09-18
• Subjects: edge detection -- support vector machines -- image texture -- image denoising; fractional‐integral‐operator‐based improved SVM -- ‐pepper noise -- image features -- texture information -- greyscale images -- support vector machine denoising method -- mapping relation -- neighbourhood signal pixels -- central pixel -- 25‐dimensional input sample -- neighbourhood pixels -- eight‐direction fractional integral operators -- centre signal pixel -- corresponding output sample -- training sample -- SVM denoising model -- centre pixel value -- noise pixel values -- final denoised image -- denoising effect -- fractional order -- visually pleasing denoised image -- clear edge information -- superior peak signal‐to‐noise ratio -- free energy -- state‐of‐the‐art denoisers; Image processing -- Periodicals; 621.36705
• DOI: 10.1049/iet-ipr.2019.0624
• Languages: English
• ISSNs: 1751-9659

Abstract

Abstract : To protect edge and texture information, when removing salt‐and‐pepper (SP) noise in grayscale images, a support vector machine (SVM) denoising method is employed. First, a mapping relation between the neighborhood signal pixels and the central pixel is designed. The size of the neighborhood is a 5 × 5 region, with a signal pixel in the center. In this region, a 25‐dimensional input sample is constructed using the correlation between the neighborhood pixels and the eight‐direction fractional integral operators. The center signal pixel acts as the corresponding output sample to provide a training sample. Then, the SVM is trained with all training samples, and the SVM denoising model is obtained. Next, the center pixel value is estimated using the SVM denoising model in every 5 × 5 region with a noise pixel in the center. Finally, the noise pixel values are replaced with the estimated values of the SVM. The experiments demonstrate that the best denoising effect is obtained when the fractional integral order is in the range of 1.8 ± 0.1. The proposed method produces a visually pleasing denoised image and obtains superior image quality assessment indicators. Our method has significant advantages compared with state‐of‐the‐art denoisers when a low level of noise is present.