A Survey on Artificial Intelligence in Pulmonary Imaging.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Pulmonary Medicine

Background:

• Deep learning (DL) has transformed computer vision and AI applications.
• Pulmonary diseases affect a significant portion of the global population.
• Medical imaging plays a crucial role in understanding lung diseases and their progression.

Purpose of the Study:

• To survey the applications of DL in pulmonary imaging.
• To highlight DL's role in diagnosing and managing lung diseases.
• To focus on DL-based segmentation of pulmonary and thoracic anatomies.

Main Methods:

• Review of DL architectures and methods applied to pulmonary imaging.
• Analysis of DL for image classification, recognition, registration, and segmentation.
• Emphasis on segmentation of lung volumes, lobes, vessels, airways, and thoracic musculoskeletal structures.

Main Results:

• DL offers significant opportunities in AI-driven medical imaging for lung diseases.
• DL techniques are effective for various pulmonary image processing tasks.
• DL-based segmentation of anatomical structures is a key application area.

Conclusions:

• DL is a powerful tool for advancing pulmonary research and clinical practice.
• Accurate segmentation of lung and thoracic anatomy using DL aids disease understanding.
• This survey provides a comprehensive overview of DL in pulmonary imaging.