An augmentation-enhanced hybrid deep learning framework for lung cancer classification using the LC15000 histopathological image dataset

Lung cancer is one of the leading causes of death from cancer around the world. Early diagnosis is necessary for effective treatment. In Bangladesh, air quality is at an alarming state, especially in the capital city of Dhaka and other major cities, where high levels of air pollution are causing respiratory problems and increasing the vulnerability to lung diseases. This has aggravated the issue of unhealthy lungs even more. Since the quality of air is already poor, there is an immediate need to discover lung cancer through histopathological images, a field in which automation and a deep learning approach can expedite the accurate process of detection. This work deals with a system for lung cancer detection and classification from histopathological images by using deep learning architecture, which employs the LC15000 dataset. The methodology uses CNNs, which is a deep learning technique, to classify histopathological images of lung cancer. Lung cancer images are preprocessed to ensure consistency and suitability for model training, with augmentation techniques applied to enhance the dataset. A carefully selected CNN architecture will be trained on a subset of the dataset and tested on validation and test sets. The model’s diagnostic capability will be assessed by important metrics like accuracy, precision, recall, and F1-score. The Hybride Model that we have developed, it’s with Augmentation: Accuracy: 0.9967, Precision: 0.9967, Recall: 0.9967, F1 Score: 0.9966 and it’s without Augmentation: Accuracy: 0.9913, Precision: 0.9913, Recall: 0.9913, F1 Score: 0.9913. It is expected that the deep learning model will be highly discriminative in classifying lung cancer from benign tissue and contribute to developing an automated diagnostic tool for pathology. This has big implications for Bangladesh, where poorly maintained air quality gives the population a higher rate of lung disease, while also facilitating the integration of AI-driven methods in a clinical setting for more efficient and accessible healthcare solutions.