Efficient Kidney Cancer Classification from CT Images Using a Lightweight Convolutional Neural Network Optimized with an Enhanced Crow Swarm Optimization Algorithm

Kidney cancer is among the fifty most common cancers globally; therefore, early and accurate classification can significantly enhance patient prognosis.However, existing classification models face challenges in effectively handling data obtained from CT imaging.This study proposes a lightweight and automated classification framework for kidney cancer detection using a hybrid feature extraction approach combined with a novel lightweight Convolutional Neural Network (CNN) optimized by a hybrid Crow Swarm Optimization (CSO) algorithm.Two datasets were used for model development and validation: the CT KIDNEY DATASET: Normal-Cyst-Tumor and Stone, containing 12,446 CT images across four classes (normal, cyst, stone, and tumor), and the Kidney Cancer Dataset with 8,400 CT images.For this study, only the normal and tumor classes were used.The feature extraction process employed multiple descriptors to derive relevant features, followed by feature selection optimized using the hybrid CSO algorithm, which demonstrated improved performance.The proposed model achieved 100% accuracy, an F1-score of 97.49%, precision of 97.97%, and recall of 98.28%, with fast processing and successful differentiation of kidney pathologies.This efficient and accurate framework, which integrates deep learning with conventional methods, presents a promising real-time kidney cancer classification system to support radiologists in clinical diagnosis and enhance detection reliability.