Analysis of the Application of Bi-Orthogonal Wavelet Transform, Bayesthresholding and Independent Component Analysis (ICA) on Poisson Noise Removal from X-Ray Images

Medical field produces huge volume of images, which are used during disease diagnosis. X-Rays are the oldest and most frequently used form of medical imaging. These images are used in many applications with prominent use found in fracture detection. The X-Ray images are normally affected by Poisson noise, which degrades the visual quality of the image and obscures important information required for accurate diagnosis. The current need is, thus, a method that removes noise while preserving important diagnostic data. This study proposed a method that combines Multiple Wavelet Denoising (MWD) Structure with ICA to remove Poisson noise from X-Ray images. The thresholding method used is Bayes Shrink and both soft, hard thresholding methods are analyzed. From the experimental results, it is evident that the proposed model produces images, which are visually clean and smooth, in fast manner. At the same time, the proposed method also preserves edges and other significant details of the image.