Innovative AI Techniques of Multiclass Classification of Liver Tumor in NIFTI Images

Abstract

Classification of Liver tumor in medical imaging causes great problems due to the complexity and variability of tumor features. This paper introduces a new deep learning system for the or the categorization of different liver diseases using NIFTI type MRI data. The system integrates learning-based 3D UNet models and Hybrid Efficient Nets into one system for accurate classification and segmentation. Advanced data preprocessing, augmentation, and the class imbalance techniques ensure standard performance. The proposed method is designed to consider tumor heterogeneity and class differences and therefore has high accuracy, sensitivity, and reliability in clinical situations. Future studies will include multivariate analyzes and patient-specific data to reach a more accurate diagnosis. This research bridges the gap between advances in mathematics and real-world clinical applications and offers solutions that can be developed to improve patient outcomes.