Lung segmentation method with dilated convolution based on VGG-16 network.

Area of Science:

• Medical Imaging
• Computer Vision
• Artificial Intelligence

Background:

• Accurate lung parenchyma segmentation in CT images is crucial for diagnosing lung diseases.
• Existing segmentation methods face challenges in precision and efficiency.

Purpose of the Study:

• To develop an accurate and effective lung parenchyma segmentation method for CT images.
• To improve the initial step in computer-assisted diagnosis of lung diseases.

Main Methods:

• A novel segmentation method combining VGG-16 network architecture and dilated convolutions.
• Utilizing the initial VGG-16 layers for image convolution and pooling.
• Employing multiple dilated convolutions to expand the network's receptive field.
• Fusing multi-scale convolutional features and using Multi-Layer Perceptron (MLP) for pixel-wise prediction.

Main Results:

• The proposed method achieved a Dice Similarity Coefficient (DSC) of 0.9867 on a dataset of 137 CT images.
• Demonstrated superior performance compared to conventional segmentation methods.
• Effectively segmented the lung parenchymal regions in CT scans.

Conclusions:

• The VGG-16 and dilated convolution-based method provides accurate lung parenchyma segmentation.
• This technique offers a significant advancement for computer-assisted lung disease diagnosis.
• The method shows potential for enhancing diagnostic accuracy in medical imaging.