A Computer Based Performance Analysis of Ultrasound Image of Liver

Abstract

In the domain of computer vision, visual analysis of human organs is currently one of the most active research issues. The key factor in designing machine capability is to do analysis of human organs. Medical imaging is the process by which physicians can evaluate an area of the human body that is not normally visible. Medical Image Processing, Analysis, and Visualization application enables quantitative analysis and visualization of medical images of numerous modalities such as Positron Emission Tomography (PET), Magnetic Resonance Imaging (MRI), Computed Tomography (CT), and Ultrasound (US). The Ultrasounds imaging is one of the main imaging modality used for diagnosis and it has advantages such as, low-cost, non-invasive, non-ionizing and being portable. newlineImage pre-processing is one of the important steps used in medical image analysis to improve the quality of an image. In order to improve the visual quality of image and to remove the unnecessary values in image, noise reduction methods are used. Image pre-processing techniques are necessary, in order to remove the speckle noise and to enhance the quality of the image. For the pre-processing, a hybrid model has been proposed based on Median and Weiner combined with Neural Network (Hybrid Neural Filter: HNF) in order to reduce the speckle noise in the ultrasound images of liver. The input image has been given to the median filter, weiner filter and diversely to the neural network. The neural network makes the comparison between the median filter output, weiner filter output and the original image, which gives the less in mean square error, that will give the output image. The output image is a filtered image. In addition to Hybrid Neural Network, Contrast-Limited Adaptive histogram equalization approach (CLAHE) is also employed as pre-processing technique, in order to enhance the visual quality of the images. newlineIn segmentation stage, two algorithms for liver tumor segmentation using ultrasound image have been proposed. In the first algorithm, Adaptively Regulariz