NanoChest-Net: A Simple Convolutional Network for Radiological Studies Classification.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Radiology

Background:

• Radiological studies like X-rays are crucial for diagnosing various diseases including COVID-19, pneumonia, and tuberculosis.
• Computer-aided diagnosis (CAD) systems enhance the analysis of radiological images.
• Current deep learning models, particularly convolutional neural networks (CNNs), require substantial data and computational power for effective disease detection.

Purpose of the Study:

• To introduce NanoChest-net, a novel, compact convolutional neural network (CNN) model.
• To evaluate NanoChest-net's efficacy in classifying multiple diseases from radiological images.
• To demonstrate a high-performing yet computationally efficient deep learning solution for medical image analysis.

Main Methods:

• Development of NanoChest-net, a lightweight CNN architecture.
• Training and validation of NanoChest-net on five diverse radiological image datasets.
• Comparative analysis against established deep learning models like ResNet50, Xception, and DenseNet121.

Main Results:

• NanoChest-net achieved superior classification performance on two of the five datasets.
• The model demonstrated comparable performance to state-of-the-art algorithms on the remaining datasets.
• NanoChest-net effectively classifies tuberculosis, pneumonia, and COVID-19 from X-ray images.

Conclusions:

• NanoChest-net offers a powerful and efficient tool for disease classification in radiological studies.
• The model's small size and reduced operational requirements make it suitable for broader clinical application.
• NanoChest-net represents a significant advancement in developing accessible AI-driven diagnostic tools for medical imaging.