Segmentation and classification on chest radiography: a systematic survey.

Area of Science:

• Medical Imaging Analysis
• Artificial Intelligence in Radiology
• Pulmonary Disorder Diagnostics

Background:

• Chest radiography (X-ray) is crucial for diagnosing lung conditions, but interpretation can be challenging.
• Computer-aided detection and diagnosis systems are essential to improve accuracy and efficiency.
• Deep learning methods now dominate medical image analysis, including chest X-rays.

Purpose of the Study:

• To review research on lung segmentation and pulmonary disorder detection/classification using chest X-rays.
• To include studies employing Generative Adversarial Network (GAN) models for segmentation and classification.
• To provide a comprehensive overview of techniques, including pre-deep learning approaches, for chest X-ray analysis.

Main Methods:

• Systematic review of research on chest X-ray analysis for pulmonary disorders.
• Inclusion of studies utilizing traditional computer vision and deep learning techniques.
• Focus on Generative Adversarial Network (GAN) applications for segmentation and classification.

Main Results:

• Deep learning, particularly GANs, is increasingly used for feature extraction, segmentation, detection, and classification in chest X-ray analysis.
• Publicly available datasets are instrumental in advancing AI-driven pulmonary disorder diagnostics.
• A historical perspective from traditional methods to deep learning provides a complete field overview.

Conclusions:

• This survey offers a dedicated resource on chest X-ray analysis techniques for pulmonary disorders.
• It highlights the evolution from traditional computer vision to advanced deep learning and GANs.
• The study aids readers in understanding current approaches and their significance in medical imaging.