Enhancing Lung Cancer Detection: Optimizing Deep Learning With Convolutional Block Attention Module

Abstract

Lung cancer manifests as the uninhibited propagation of cells within tissues in lung. Early exposure of malignant cells within the lungs, which play crucial roles in oxygenation and carbon dioxide exchange vital for human physiology, is imperative. Deep learning in detecting lung cancer has gained considerable attention because of its impending to enhance patient identification of diagnosis and treatment. Nonetheless, existing algorithms demonstrate less-than-ideal performance in crucial areas such as recognition accuracy, precision, sensitivity, F-score, and specificity. To address this issue, a novel method Unet with CBAM is leveraged by the different deep learning pre trained models such as VGG-16, VGG-19, ResNet-50, Resnet-101 and EfficientNet. In this proposed methodology we introduce Unet with Convolutional Block Attention Module (CBAM) for the prediction of lung cancer. The proposed model has a base learner consists of two modules: Unet and CBAM attention. In the attention mechanism there are also two modules: channel attention module and spatial attention modules. The features of each module are integrated and supplied to the meta-learner as a fully connected model for the precision. This model improves results comparison with other models.