ARaNN: An Adaptive Snake-Optimized Associated Random Neural Network for Liver Cancer Detection and Classification in CT Images

Abstract

 Liver is the largest organ of our body and the cancer that forms inside the liver cell is called Liver Cancer (LC). Many types of liver cancers are there and the most common type is known as Hepatocellular carcinoma which starts with the cell that present in the liver hepatocyte. Less common LC types are hepatoblastoma and intrahepatic cholangiocarcinoma. The detection and classification of LC is an arduous process and performing it by manually is a time overwhelming and imprecise task. Hence now-a-days artificial intelligence based approach has been used by many hospitals. In context with this, we propose a novel optimized deep learning approach known as Adaptive Snake optimization algorithm (ASA) based Associated Random Neural Network approach (ARaNN) to identify and detect LC. To begin the process, the CT scan images are pre-processed to ignore the presence of noise and to improve the contrast to predict the features effectively. Henceforth, Fuzzy Archery Learning (FAL) algorithm is applied to segment the images followed by extracting the features and detection and classification of LC using the proposed approach. Simulations are effectuated to authenticate the performance of the proposed effort. The outcomes are analyzed by the analogous study with state-of-art works and our approach surpasses all the other approaches while detecting the LC.