Liver is a vital organ to detoxify the blood and break down the fat molecules from the food that we consume. Most of the liver diseases are inherited from their parents. Liver can get damaged due to various factors, the liver can be damaged due to many reasons. Among this, cancer is one cause of liver damage. Considering the Indian population, the rate of mortality due to liver cancer is 6.8 per one lakh in men and 5.1 per one lakh in women. The computed tomography is the common method to screen, diagnose, staging, and prognosis assessment; thereby, the liver cancer patient can be given proper treatment. While using image processing techniques for liver tumor segmentation, we adopted a binary segmentation task where a network has been designed to segment liver and liver tumor segments. The proposed system focuses on medical image sequences of the liver tumor segmentation based on region. The abnormality structure of the liver has been altered based on the position and by setting a threshold for segmenting. A deep neural network was used to compare the dataset images, and the input dataset has been trained and verified. Finally, the status of liver tumor would be inferred whether it is malignant or benign. © 2023 Elsevier B.V., All rights reserved.