Deep Convolution Neural Network in classification of liver tumor as benign or Malignant from Abdominal Computed Tomography
S. Gunasundari, S. Meenambal, Selvaraj Tamilselvi, R Dhanalakshmi · 2022 Third International Conference on Intelligent Computing Instrumentation and Control Technologies (ICICICT) · 2022
The most shared reason of cancer demise is liver cancer, which is detected by analysing changes in the grey level liver tissue in computer Tomography images. A computer-assisted diagnostic approach for characterization of liver tumours from abdominal Computed Tomography (CT) is proposed in this research to assist doctors in reducing misdiagnosis. The objective of this study is to generate a computer-aided diagnostic model that uses a deep Convolutional Neural Network (CNN) to differentiate between benevolent and malevolent liver tumours. With the help of a clinician, ROIs belonging to focal nodular hyperplasia, hemangioma, hepatoma, and cholangiocarcinoma are segmented. The features are extracted and decision model is built using CNN. The developed CNN model can able to differentiate liver diseases such as focal nodular hyperplasia and hemangioma as benign, hepatoma and cholangiocarcinoma as malignant. The result shows that the developed CNN model gives very good accuracy of 91%.