Lung Cancer Detection using CNN and SVM

Abstract

Lung cancer is one of the main causes of cancer-related mortality worldwide. Early diagnosis is vital to improving treatment outcomes. Medical imaging benefits greatly from deep learning, a branch of artificial intelligence that uses hierarchical neural networks to automatically extract and learn characteristics from massive datasets. This study introduces a deep- learning approach that is hybrid and uses Convolutional Neural Networks (CNNs) for feature extraction and Support Vector Machines (SVMs) to identify lung cancer subtypes from CT images. We trained a CNN algorithm to extract robust features by preprocessing the CT scans using a well-structured dataset from Kaggle that covers both benign and malignant lung diseases. To increase accuracy and generalization, the collected characteristics were then fed into an SVM classifier. The model surpassed traditional approaches in terms of speed and predictive power, as measured by accuracy (93.2%), F1- score (91.7%), and specificity (94.5%). Data augmentation methods were also applied to increase the resilience of the model. The suggested technique shows a great deal of promise for helping radiologists diagnose lung cancer early.