Grayscale To Colour Enhancement Of Liver Ct Scan Images And Classification Of The Hepatic Lesions

Abstract

Liver diseases are varied and most of them involve morbidity and mortality. The most common of them are the hepatic lesions which are of multiple types and each require different methods of diagnosis and treatment. Computerised Tomography (CT) is the most common modality for detection of liver lesions.The CT image is in grayscale hence the internal differences are difficult to detect and so the diagnosis becomes complex. To remove this difficulty the image has been converted to an image with contrasting colours. This improves its sensitivity and specificity. The image has been initially preprocessed to remove noise and other non-linearities and then enhanced from grayscale to colour. In the image if a lesion is identified in the liver then that region alone is extracted using the ROI segmentation. It is then classified as benign(non-cancerous) or malignant(cancerous) [using the SVM classifier]. This is done since each of the lesion types require different methods of treatment, surgical or non-surgical.