Effective Segmentation of Liver CT images using Marker Controlled Watershed Algorithm

Many important tasks such as blood purification and removing toxic elements from body are performed by liver which makes it a very vital organ in human body. Changing lifestyle has increased the risk of liver cancer in recent times. Traditional methods of liver cancer detection take more time and it is costly. Therefore, computer aided diagnosis has gained popularity due to their ability to detect cancer in less time along with less costly. Computer aided diagnosis involves processing computed tomography images using some machine learning algorithms or deep learning algorithms. Efficiency of these algorithms depend upon the quality and quantity of input images. Noises are inherent in medical images. Noise can be removed using some pre-processing techniques. Another important step is segmentation, which involves separating unwanted organs from medical images to obtain region of interest. This article presents an effective segmentation technique using Marker Controlled Watershed Algorithm. To evaluate the effectiveness of proposed method dice score, volume overlapping error, relative volume difference, and Jaccard Index are used as evaluation parameters.