期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2018
卷号：7
期号：3
页码：328-334
出版社：Shri Pannalal Research Institute of Technolgy
摘要：All medical images are suffering from some kind of noise. Low contrast, blurring, and poor quality are vital problems that appear in the production of medical images. OPG (orthopantomography) images are also suffering from the same above problems; so that many types of filtration techniques have been proposed to overcome with these drawbacks because noise can affect the quality and validation of diseases diagnosis. This paper proposes a hybrid method of noise filtration for OPG images based on adaptive median filter and discrete wavelet transform. The analysis of simulation results show that the proposed method produces more desirable results for mixed noise denoising which contains Salt & Pepper, Gaussian, and Speckle noise compared with three denoising approaches which are standard median filter, adaptive median filter, and discrete wavelet transform. The proposed method achieves averaged PSNR of 31.38 dB and averaged SSIM of 0.9042 for mixed noise.
关键词：Image denoising; OPG image; Adaptive median filter; Discrete wavelet transform.
