SAP 잡음 제거를 위한 적응적 스위칭 필터링 알고리즘

• Published in: (사)디지털산업정보학회 논문지, 18(1) Vol. 18; no. 1; pp. 25 - 35
• Main Authors: 김동형, Kim, Donghyung
• Format: Journal Article
• Language: Korean
• Published: (사)디지털산업정보학회 01.03.2022
• Subjects: 컴퓨터학; ASF(adaptive switching filter); NAMF(non-local adaptive mean filter); SAP Noise Reduction; AFSWMF(adaptive fuzzy switching weighted mean filter); RWMF(reliability-based mean filter)
• ISSN: 1738-6667; 2713-9018

The SAP(salt-and-pepper) noise changes the pixel value to the maximum and minimum values of the dynamic region of the pixel. For this reason, unlike white Gaussian noise, SAP noise can predict the ratio of noise relatively easily. Because the condition of the neighboring pixels that can be referenced changes according to the noise ratio, it is necessary to apply different noise reduction methods according to the noise ratio. This paper proposes an adaptive switching filtering algorithm which can eliminates the SAP noise. It consists of two phases. It first detects the location of the SAP noise and calculates the noise ratio. After that, the image is reconstructed using different methods depending on which of the three sections the calculated noise ratio belongs to. As a result of the experiment, the proposed method showed superior objective and subjective image quality compared to the previous methods such as MF, AFSWMF, NAMF and RWMF.