FUZZY ENHANCED KIDNEY TUMOR DETECTION INTEGRATING MACHINE LEARNING OPERATIONS FOR A FUSION OF TWIN TRANSFERABLE NETWORK AND WEIGHTED ENSEMBLE ML CLASSIFIER

Abstract

This study addresses the critical challenge of kidney tumor detection from CT scans by proposing a novel framework called "Fuzzy Enhanced Kidney Tumor Detection." The approach incorporates fuzzy image enhancement techniques to improve the quality and contrast of kidney CT images, thereby facilitating more accurate tumor detection. This preprocessing step is crucial for reducing noise and highlighting tumor boundaries effectively, allowing downstream models to better differentiate between healthy and tumorous tissues.The core innovation lies in integrating machine learning operations through a fusion of a Twin Transferable Network and a Weighted Ensemble Machine Learning Classifier to perform tumor classification. The Twin Transferable Network is designed to leverage transferable features across different datasets or imaging conditions, enhancing the model's generalization capability. Meanwhile, the Weighted Ensemble classifier combines predictions from multiple machine learning algorithms, weighted by their individual performance, to produce a robust and reliable final classification. This hybrid approach not only capitalizes on the strengths of various classifiers but also mitigates their weaknesses, resulting in higher accuracy and robustness in kidney tumor detection. The proposed system holds significant potential for aiding radiologists and improving diagnostic workflows in clinical settings. This work contributes to the advancement of AIdriven medical imaging by offering an effective combination of fuzzy image processing and hybrid machine learning architectures tailored for kidney tumor diagnostics. Experimental results demonstrate its superiority over traditional methods, reinforcing its potential as a valuable tool in computer-aided diagnosis.