Due to the vast variety of human differences in the morphologies of the liver and the variance in pixel intensity in the picture, automatic liver segmentation is challenging. Furthermore, the limits of the liver are unclear since it shares intensity distributions with neighbouring organs and tissues. We suggest a quick and accurate approach for segmenting the liver using contrast-enhanced computed tomography (CT) images in this methodology. We apply level-set speed photos to adopt the two-step seeded region growth (SRG) method to generate an initial liver boundary that is roughly defined. According to the gradient information and related component connectivity, this separates a CT picture into a collection of distinct objects. Our technique reduces computing time by reducing threshold propagation, which converges at the best segmentation result, with such an optimal estimation of the initial liver border. The limits of liver computer supported detection/diagnosis systems and potential strategies to enhance them will also be covered. We came to the conclusion in this research that, despite some potential for advancement, automatic liver segmentation approaches are now on level with human segmentation. However, it may be said that both automated and semi-automatic liver tumour segmentation approaches perform less well than anticipated. It is also clear that the majority of computer assisted detection/diagnosis techniques call for manual liver and liver tumour segmentation, which restricts the clinical use of these systems.