Design and Development of Medical Image Processing Techniques and to Study Their Application using Graphical System Design in Ovarian Cancer

Abstract

Ovary is a female reproductive organ located on each side of the uterus that newlinestore eggs. Ovaries have two reproductive glands that produce eggs. Ovary produces newlinefemale hormones such as estrogen and progesterone. Cancer develops from this organ newlinewhen abnormal cells grow uncontrollably. The growth of the abnormal cells in the newlineorgan develop tumor and creates pressure on the nearby organs. Ovarian cancer is a newlinedeadly disease for which there is no effective means of early detection. Early newlinediagnosed and treated of ovarian cancer will improve survival rate. Ovarian newlinecarcinomas comprise a diverse group of neoplasms, exhibiting a wide range of newlinemorphological characteristics, clinical manifestations, genetic alterations, and tumor newlinebehaviors. This high degree of heterogeneity presents a major clinical challenge in newlineboth diagnosing and treating ovarian cancer. Early detection of ovarian cancer is newlinedifficult because of its unknown symptoms. There are no specific symptoms for newlineovarian cancer but the people may experience abdominal bloating, change in bowel newlinehabits, fluid in the abdomen, indigestion and nausea. Ultrasound images are taken to newlinedetect the abnormality. In this research work ultrasound images of the ovary are used. newlineThe novel method to detect the ovarian cancer is proposed. This method is composed newlineof four main processes. First, the image is preprocessed using the wiener filter. newlineFiltered images are enhanced using contrast stretching. Stretched image is used for newlinesegmentation by using three clustering techniques. Second, clustering techniques such newlineas K-means clustering, FCM and Adaptive clustering are used to cluster the image. newlineTrained set of images are used in this research work. Feature extraction is done from newlinethe trained set of images using GLCM technique. Third, GLCM feature extraction newlinetechnique is used to extract the features present in the image. Feature extracted images newlineare classified using support vector machine technique. Fourth, support vector machine newlinetechnique is used to classify the normal and abnormal images. Fifth, comparative newlineanalyses for the three algorithms are done using four parameters such as F-Score, newlineDice, Recall and precision. Performance analyses for the three algorithms are newlinemeasured using accuracy, sensitivity, specificity, time period and iteration. Finally, newlineFuzzy c-means clustering algorithm produce the best result when compared to other newlinetwo algorithms by means of comparative graphs and parameters. newline