An Efficient Deep Transfer Learning based Framework for Accurate Liver Disease Prediction

Abstract

Liver disease poses a significant global health burden, leading to millions of deaths annually due to late diagnosis and inaccurate classification. Early and precise identification of liver abnormalities such as ballooning, fibrosis, inflammation, and steatosis is crucial for effective treatment and improved patient outcomes. Traditional diagnostic methods rely heavily on manual histopathological examination, which is both time-consuming and prone to human error. To address these limitations, this study proposes an Efficient Deep Transfer Learning-based Framework for accurate liver disease prediction using histopathological images. The proposed framework integrates advanced convolutional neural network (CNN) architectures and transfer learning techniques to automatically extract and classify disease-specific features, thereby enhancing diagnostic precision and reducing inter-observer variability. The methodology involves collecting a large dataset of verified liver tissue images, followed by preprocessing steps such as normalization, resizing, noise reduction, and augmentation to improve data quality and model generalization. Several pre-trained deep learning models EfficientNetB2, DenseNet121, InceptionV3, ResNet50, and VGG16 are fine-tuned using transfer learning to optimize classification accuracy. The models are evaluated using metrics such as accuracy, precision, recall, F1-score, and AUC-ROC to ensure robust performance. Experimental results demonstrate that the proposed framework achieves superior accuracy and consistency compared to conventional methods.