Efficient Technique for Removal of White and Mixed Noises in Gray Scale Images
16 Pages Posted: 7 Sep 2019
Date Written: September 3, 2019
Images are often affected by different kinds of noise while acquiring, storing and transmitting it. Even the datasets gathered by the various image acquiring devices would be contaminated by noise. Hence, there is a need for noise reduction in the image, often called Image De-noising and thereby it becomes the significant concerns and fundamental step in the area of image processing. During image de-noising, the big challenge before the researchers is removing noise from the original image in such a way that most significant properties like edges, lines, etc., of the image should be preserved. There were various published algorithms and techniques to de-noise the image and every single approach has its own limitations, benefits, and assumptions. This paper reviews the noise models and presents a comparative analysis of various de-noising filters that involve in producing a high-quality image. The metrics like PSNR (Peak Signal to Noise Ratio), Entropy, SSIM (Structured Similarity Index), MSE (Mean Squared Error), FSIM (Feature Similarity Index), and EPI (Edge Preserving Index) are considered as image quality assessment metrics.
Keywords: De-noising, Edge preserving filtering, Spatial Domain Filters, Transform Domain Filters, Non Local Means, DnCNN, Gaussian noise, Mixed Noise, PSNR, MSE, EPI, FSIM, SSIM
Suggested Citation: Suggested Citation