Lung Nodule Classification using A Novel Two-stage Convolutional Neural Networks Structure'.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Lung cancer remains a leading cause of cancer-related mortality globally.
• Early diagnosis significantly improves patient survival rates after treatment.
• Computed Tomography (CT) imaging is a vital tool for lung cancer detection.

Purpose of the Study:

• To introduce a novel two-stage convolution neural network (2S-CNN) for lung cancer classification from CT images.
• To enhance the accuracy of lung cancer diagnosis through advanced image analysis.
• To refine ambiguous CT images and improve recognition of complex patterns.

Main Methods:

• Development of a two-stage Convolution Neural Network (2S-CNN) architecture.
• The first stage employs a basic CNN to refine input CT images and extract ambiguous regions.
• The second stage utilizes a simplified Inception CNN (GoogLeNet variant) for enhanced recognition of complex features.

Main Results:

• The proposed 2S-CNN model achieved a classification accuracy of 89.6% on lung CT images.
• The two-stage approach demonstrated effectiveness in refining images and recognizing complex patterns.
• The system successfully identified features indicative of lung cancer in CT scans.

Conclusions:

• The novel 2S-CNN model shows significant promise for accurate lung cancer classification from CT images.
• Early detection of lung cancer can be enhanced through advanced deep learning techniques.
• This approach offers a potential improvement in diagnostic tools for radiologists.