To introduce ERYXSeg, a deep learning architecture for accurate and efficient wound segmentation.
Approach:
Key Findings:
ERYXSeg achieves the highest reported IoU and Dice scores: 0.7633 IoU and 0.8658 Dice for the foot ulcer dataset; 0.6910 IoU and 0.8173 Dice for the curated mixed wound dataset.
Attention-gated skip connections are crucial for accurate spatial reconstruction.
Mobile Inverted Bottleneck Convolution (MBConv) blocks are essential for feature extraction.
Interpretation:
The results demonstrate excellent generalization and high computational efficiency.
Limitations:
The model's performance may still be influenced by the variability in wound characteristics and imaging conditions.
Limited availability of annotated datasets may affect the robustness of the model.
Conclusion:
ERYXSeg represents a significant advancement in automated wound segmentation, addressing key challenges in the field.
