A novel method based on Wiener filter for denoising Poisson noise from medical X-Ray images. (January 2023)

• Record Type: Journal Article
• Authors: Göreke, Volkan
• Is Part Of: Biomedical signal processing and control. Volume 79(2023)Part 1
• Publication Date: 2023-01
• Subjects: Poisson noise -- X-ray image filter -- Artificial intelligence; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2022.104031
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: A novel method was implemented within a nested architecture where a 2D FIR filter algorithm was embedded in the standard Wiener algorithm. Common optimization algorithms were tested individually to obtain the 2D FIR coefficients. Poisson noise is superbly filtered in both synthetic and medical images. Abstract: Background and Objective: The Poisson noise is added to the image during the acquisition of medical X-Ray images. The distorted image due to this noise makes it difficult for physicians to diagnose the disease. Although there are various approaches for filtering Poisson noise, these approaches have disadvantages such as excessive smoothing of the image, distorting the texture information, reducing the image quality and high computational cost. In this study, a novel method that removes Poisson noise from medical X-Ray images is proposed by overcoming the above mentioned disadvantages. Methods: In the proposed method, the Wiener filter is modified using the FIR filter embedded in the standard Wiener algorithm. The FIR filter design was carried out using the ASO optimization algorithm. Optimum local mean and optimum local variance values are calculated using the optimization matrix corresponding to the FIR filter coefficients and transferred to the standart Wiener filter layer as parameter inputs. Results: The proposed method showed superior performance in synthetic images and medical X-Ray images in terms of PSNR, MSE, SSIM metrics and image quality metrics such as luminous intensity, Contrast index, Entropy and Sharpness. The time consumption of the proposed method is much less. Conclusions: The clinical usage of the proposed method may help doctors to be able to diagnose the disease more accurately by interpreting the X-ray images. Besides, the proposed method can also have a positive effect on the CAD performance by using it at the pre-processing stage of CAD systems.