Liver disease is a serious health condition that affects millions of individuals worldwide. Radiomics and deep learning (DL) are two AI-based methods that are being actively researched in the field of liver imaging. The analysis of liver related problems is a tedious tasks and it requires models to classify and predict. Radiomics is an effective method for extracting an extensive amount of distinctive characteristics from medical images using data-characterization techniques. Radiomic features possess the ability to reveal tumoral patterns and traits that the human eye alone cannot detect. Convolutional Neural Networks (CNN) is well suited method to analyze visual data and it is most useful to diagnose the liver diseases with extracted radiomics features. This study suggests an extraction of radiomics features and classifies it using CNN for liver diseases. The combination of Radiomics and CNN improves the prediction performance. © 2025 Elsevier B.V., All rights reserved.