Gardado en:
| Autor Principal: | |
|---|---|
| Formato: | Recurso digital |
| Idioma: | inglés |
| Publicado: |
Zenodo
2025
|
| Subjects: | |
| Acceso en liña: | https://doi.org/10.5281/zenodo.15499500 |
| Tags: |
Engadir etiqueta
Sen Etiquetas, Sexa o primeiro en etiquetar este rexistro!
|
Table of Contents:
- <p dir="ltr">Lung cancer stands as the foremost cause of cancer-related mortality in the United States, claiming approximately 127,000 lives annually. While existing efforts in lung cancer detection demonstrate effectiveness, none comprehensively address the entire spectrum of issues associated with this condition. In response to this challenge, we developed the Cognitive Cyber Lung (CCL) AI, a multi-stage program designed with the overarching objectives of 1) identifying lung cancer, 2) localizing cancerous nodules within the lungs, 3) providing actionable feedback for surgeons, and 4) enhancing surgical understanding through nodule visualization.</p> <p dir="ltr">The creation of CCL involved leveraging PyTorch, a robust Python-based AI programming language, for model development. Employing advanced techniques such as image segmentation, Res-Net 101 architecture, and U-net architecture, we tailored Convolutional Neural Network (CNN) models to address the intricacies of lung cancer identification and nodule localization. These models consistently achieved average accuracies of 70% or higher across all specified objectives. This research represents a noteworthy advancement in lung cancer diagnostics and surgical guidance. The CCL AI model contributes a robust tool that comprehensively aids medical practitioners throughout the entire spectrum of lung cancer management, underscoring the potential of our approach to opening avenues for more precise and informed decision-making in the ongoing battle against lung cancer. </p> <p><strong><br>Keywords: Lung Cancer, Artificial Intelligence, Machine Learning, Deep Learning, Convolutional Neural Networks, PyTorch, Google TPU, Surgical Diagnostics</strong></p>