SparseCNNet: An Efficient Deep Learning Architecture for Rapid Lung Cancer Detection

Lung cancer is one of the cancers that makes most of the cancer-related deaths worldwide, with early detection being critical to improving survival rates. Computed Tomography (CT) imaging is one of the primary tools used in identifying and diagnosing lung cancer due to its high-resolution capabilities. However, manual analysis of CT images can be time-consuming and may lead to errors. This project proposes a Convolutional Neural Network (CNN)-based model designed to classify the images into whether they are benign, malignant or normal lung CT scans. By using Sparse Pooling technique in this CNN architecture, we aim to reduce computational complexity while retaining model accuracy, making it for real-time implementation. This below model is trained and validated on the IQ-OTHNCCD Lung Cancer dataset, and the techniques utilisied are image augmentation that are used to improve robustness and reduce the chances of overfitting. The experimental results suggest that the model provides reliable predictions across various types of lung conditions, demonstrating potentially helping the radiologists in early detection of cancer and diagnosis.