Early Lung Cancer Detection Using CNN

This paper presents a unique automated strategy for lung cancer detection that employs Convolutional Neural Networks (CNNs), addressing the global health concern of early detection for improved treatment effectiveness. The study focuses on preparing a heterogeneous dataset that includes both cancerous and non-cancerous lung pictures rigorously split into training, validation, and test sets. Augmentation techniques including zooming, shearing, flipping, and normalization improve the model’s capacity to identify significant features in the mixed dataset. The CNN architecture built using Keras comprises convolutional, maxpooling, and dropout layers designed to detect detailed patterns within lung images. To avoid overfitting, early halting and model checkpointing procedures are used during training, assuring optimal performance and generalization. The trained CNN model achieves an impressive 95% accuracy on the given dataset, demonstrating its efficiency in detecting lung cancer. This study significantly advances automated lung cancer detection methods by proposing a new mixed dataset technique. The high accuracy gained demonstrates CNNs’ potential to revolutionize lung cancer detection, highlighting the critical significance of deep learning in healthcare applications. This unique methodology shows potential for assisting doctors in early diagnosis, improving patient health outcomes.