FFAE-UNet: An Efficient Pear Leaf Disease Segmentation Network Based on U-Shaped Architecture.

Area of Science:

• Agricultural Science
• Computer Vision
• Plant Pathology

Background:

• Precise segmentation of pear leaf diseases is vital for smart agriculture.
• Existing methods struggle with small disease areas, blurred edges, and background noise.

Purpose of the Study:

• To propose an improved U-Net architecture (FFAE-UNet) for accurate pear leaf disease segmentation.
• To enhance the segmentation performance by addressing limitations of current approaches.

Main Methods:

• Introduction of the Attention Guidance Module (AGM) to reduce background noise and capture feature relationships.
• Integration of the Feature Enhancement Supplementation Module (FESM) to improve responsiveness to multi-scale disease features.
• Development of the FFAE-UNet architecture combining AGM and FESM.

Main Results:

• FFAE-UNet achieved 86.60% MIoU, 92.58% Dice coefficient, and 91.85% MPA.
• The proposed model significantly outperformed existing mainstream segmentation methods.
• Demonstrated effectiveness in handling challenges like small disease areas and blurred edges.

Conclusions:

• FFAE-UNet offers a robust solution for precise pear leaf disease segmentation.
• The model aids farmers and experts in disease evaluation and management for smart agriculture.
• Improved segmentation facilitates effective pest control and disease management strategies.