Chest X-ray dataset and ground truth for lung segmentation

Area of Science:

• Medical Imaging and Artificial Intelligence
• Radiology and Computer-Aided Diagnosis

Background:

• Chest X-rays are vital for diagnosing Covid-19, with computer-assisted analysis enhancing accuracy.
• Image segmentation is crucial for early disease detection and precise diagnosis in medical imaging.
• Deep learning models require validated datasets for effective training in image segmentation and classification tasks.

Purpose of the Study:

• To present a meticulously verified dataset of chest X-ray images for AI-driven medical diagnostics.
• To facilitate the development of advanced deep learning models for disease detection and classification.

Main Methods:

• Compilation of 292 chest X-ray images from Airlangga University Hospital.
• Inclusion of radiologist-verified ground truth data for each image.
• Dataset categorizes images into Covid-19, pneumonia, and normal conditions.

Main Results:

• A novel dataset of 292 chest X-ray images with verified ground truth is now available.
• The dataset supports AI model training for segmentation and classification of lung conditions.
• Potential for improved accuracy in diagnosing Covid-19 and pneumonia using AI.

Conclusions:

• The presented dataset is valuable for advancing AI in medical image analysis.
• It enables the training of robust deep learning models for chest X-ray interpretation.
• Facilitates more accurate and efficient diagnostic outcomes in clinical settings.