Lung Cancer Detection and Classification

The research proposes a novel Cat Swarm Optimization with Deep Ensemble Learning for Lung Cancer Classification (CSODEL-LCC) technique using CT scans. The proposed CSODEL-LCC method focuses on designing a Deep Learning (DL) algorithm for the detection and classification of lung tumors from CT images. To achieve this, the CSODEL-LCC technique initially performs a segmentation process using the UNet++ algorithm, while the Cat Swarm Optimization (CSO) system conducts hyperparameter tuning. Meanwhile, a fusion-based feature extraction process is employed using the InceptionResNetV2 and EfficientNetB5 models. Finally, an ensemble classification is performed using three DL classifiers: Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Bidirectional LSTM (Bi-LSTM). The RMSProp optimizer is used to fine-tune the hyperparameters of these three DL models. The performance of the CSODEL-LCC method is validated using a benchmark CT image dataset.