A Combined Ensemble Model (CEM) for a Liver Cancer Detection System

Abstract

The liver is one of the most important organs in the human body. The liver's proper function is critical for overall health, and liver diseases or disorders can have serious consequences. Liver cancer is also known as hepatic cancer, which is divided into various types of cells that belong to the cancer. The most common type of liver cancer is hepatocellular carcinoma (HCC). HCC is one of the most common types of liver cancer that can affect up to 85% of people worldwide. Early detection of liver cancer is essential in healthcare because it increases the chances of successful treatment and patient outcomes. Many researchers have developed models that help detect and diagnose liver cancer. The first step in detecting liver cancer is identifying people at a higher risk. Chronic hepatitis B or C infection, cirrhosis, heavy alcohol use, obesity, and exposure to certain chemicals and toxins are all risk factors. This paper is mainly focused on detecting the cancer-affected regions that occur in the liver. In this paper, a combined ensemble model (CEM) for a liver cancer detection system is developed to find and detect liver cancer and liver disorders in their early stages. A pre-trained model, RESNET50 with transfer learning, is used to obtain the features from the pre-trained model-an advanced preprocessing technique involved in filtering the noise from input CT scan images. A hybrid feature extraction (HFE) technique also gets significant elements from the input CT scan images. Finally, the proposed CEM combines an Extreme Gradient Boosting (EGB) algorithm with a Recurrent Neural Network (RNN) that focuses on detecting the abnormal cancer cells present in input CT scan images. The performance of the CEM shows a high accuracy of 98.48% with a 10% high detection rate. Previously, it was 88.12%.