Liver Disease Classification using Optimized Periodic Implicit Generative Adversarial Network

Manjunath Varchagall

doi:10.1016/j.knosys.2025.114579

Liver Disease Classification using Optimized Periodic Implicit Generative Adversarial Network

2025-10-19

0

PRE

AI

DOI：10.1016/j.knosys.2025.114579

Original

Original request for help

Save

Abstract

En 中文

A healthy liver produces a significant amount of organic components in humans. But an unhealthy liver can be life-threatening and can lead to death. Only early disease detection can increase the rate of survival. Therefore, Periodic Implicit Generative Adversarial Network optimized with Golden Jackal Optimization for Liver Disease Classification (PIGAN-GJOA-LDC) is proposed in this manuscript to classify the liver disease. Initially, the input images are amassed from Indian Liver Patient Dataset (ILPD). Then the input images are given to the preprocessing stage, where Colour Wiener filtering is used to increase the quality of input image. The pre-processing images are fed to the Improved Non-subsampled Shearlet Transform (INSST) approach for feature extraction. The extracted features are supplied to the Periodic Implicit Generative Adversarial Networks (PIGAN) for classifying the liver disease and non-liver disease. The Golden Jackal Optimization algorithm (GJOA) is considered to optimize the weight parameters of PIGAN. The efficiency of the PIGAN-GJOA-LDC approach is evaluated by some metrics: accuracy, precision, f-scores. The proposed approach attains 99.87% higher accuracy which is 27.32%, 29.83%, 25.29% higher for liver disease and lower computation time of 33s which is 20.26%, 20%, 21.74% lower compared to the existing models.

Liver Disease Classification using Optimized Periodic Implicit Generative Adversarial Network

Abstract

Journal

Organization

Cited Papers