Semantic-Attention Enhanced DSC-Transformer for Lymph Node Ultrasound Classification and Remote Diagnostics

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computer Vision

Background:

• Accurate lymph node ultrasound image classification is crucial for disease diagnosis.
• Existing methods face challenges with noise and identifying diagnostically significant regions.
• The need for efficient AI models in remote diagnostic settings is growing.

Purpose of the Study:

• To introduce a novel Semantic-Attention Enhanced Dynamic Swin Convolutional Block Attention Module (CBAM) Transformer (DSC-Transformer) for lymph node ultrasound image classification.
• To improve the efficiency and accuracy of AI-driven medical image analysis, particularly for telemedicine.
• To develop a model capable of handling noise and focusing on critical diagnostic features.

Main Methods:

• Integration of semantic feature extraction with a Swin Transformer architecture.
• Implementation of multi-scale attention mechanisms (CBAM) for capturing global and local image details.
• Development of semantic-driven preprocessing and adaptive compression techniques.

Main Results:

• The DSC-Transformer demonstrated superior classification performance on diverse lymph node ultrasound datasets.
• Grad-Channel Attention Module (CAM) visualizations confirmed effective focus on diagnostically relevant areas.
• The model maintained high efficiency, suitable for remote diagnostic and telemedicine scenarios.

Conclusions:

• The DSC-Transformer offers a significant advancement in AI-driven medical image analysis for lymph node classification.
• Its semantic-attention enhancement makes it highly effective for telemedicine and remote diagnostic applications.
• The model's ability to process images efficiently while suppressing noise holds broad implications for telehealth deployment.