DP-UNet: Dual branch attention multi-layer encoder and progressive fused pyramid pooling network for COVID-19 infection region segmentation

Key Points

• Segmentation accuracy improves significantly due to the dual branch architecture and attention mechanism, enhancing detail capture.
• The method achieves a notable accuracy of 95.2% on test images, validating its efficacy in clinical scenarios.
• This analysis utilizes a novel dual branch attention multi-layer encoder approach for precise image segmentation.
• The model empowers enhanced detection capabilities in clinical settings, showcasing its potential impact on patient management.