An accurate medical digital diagnostic system using non linear filter based image quality enhancement techniques

Abstract

Medical imaging is the process of creating images of the human newlinebody for diagnosis of disease at an early stage. Several medical imaging newlinemodalities such as Magnetic Resonance Imaging (MRI), Ultrasound (US) newlineimaging, Single Photon Emission Computed Tomography (SPECT) and newlinePositron Emission Tomography (PET) are used to create the image. Among newlinethese modalities, US imaging has been widely exploited for diagnosis due to newlineits real time measurement, cost effectiveness, safety features, non-invasive newlinenature and portability. However, the drawback of US imaging is the poor newlineimage quality due to speckle noise. Speckle noise is an inherent property of newlinemedical US imaging and because of this noise the image resolution and newlinecontrast become reduced, which affects the diagnostic value of this imaging newlinemodality. It also affects the success of other processes including newlinesegmentation, feature extraction, feature reduction and recognition. Image despeckling and segmentation are the very important preprocessing steps in medical image processing. Several image denoising and segmentation methods are reported in the literature that perform well for some conditions. But each method has their own merits and demerits. This fact motivated that there is sufficient scope to develop an efficient method to improve the quality of US images. The image enhancement approaches adopted in this thesis are summarized below: Image denoising using Integer Wavelet Transform (IWT) and fuzzy filter. Image despeckling employing fractional calculus and Image segmentation based on fuzzy rules and clustering approach newline newline