Improved Spatially Adaptive Thresholding Technique for Noise Reduction

Abstract

Noise suppression in images is a difficult task and anomalous. The selection between the preservation of actual image quality and noise reduction elements must be made in a way that it enhances the diagnostically applicable image content. For example in the medical images, specialists generally do not have the biomedical expertise to judge the demonstrative importance of the De-noising results. For instance, in ultrasound pictures, dot clamor may contain data valuable to medicinal specialists that the spotted composition makes troublesome for an analysis. Similarly, Gaussian noise in an image captured by a camera at the remote or local server may have noise due to signal attenuation, dust or the poor exposure of the image to light. Also, such pictures show compelling variability and it is important to optimize with reduced CPU overhead. This persuades the development of efficient denoising methods and robust that is appropriate to different circumstances, instead of being ideal under particular conditions. In this thesis, we have proposed robust technique that adjusts to different levels of Gaussian image noise and detects it using spatially adaptive soft thresholding technique. A Gaussian deviation in X direction parameter is used to adjust the preservation of pertinent points of interest of the image against the level of noise reduction. A demonstration of its helpfulness for denoising and upgrade of the pictures is done in MATLAB environment with the addition of Gaussian noise.