Enhanced Chest Cancer Classification using Convolutional Neural Networks on CT Scan Images for Improved Diagnostic Accuracy

This work investigates employing convolutional neural networks (CNNs) for CT scan image classification of chest cancer with an eye towards higher diagnosis accuracy and efficiency. The paper classifications images into squamous cell carcinoma, adenocarcinoma, big cell carcinoma, and normal tissues using a CNN model taught and evaluated on a large dataset. Especially adept at identifying squamous cell cancer and adenocarcinoma, the model achieved an overall accuracy of $86 \%$ with excellent precision, recall, and F1-scores. Training and validation accuracies exhibited fast initial improvement, stabilizing around $90 \%$, therefore indicating excellent learning and generalization with minimum overfitting. Although the model showed good general performance, its recall for large cell carcinoma revealed opportunities for development, implying the necessity of additional optimization and a more varied dataset to increase resilience. This work emphasizes CNNs’ great promise in medical imaging and presents a viable approach for early, accurate chest cancer identification. The results highlight the relevance of deep learning in contemporary healthcare and open the path for next studies aimed on improving AI-based diagnostic tools, thereby contributing to better patient outcomes and more effective clinical diagnosis processes.