Abstract
Diabetic Foot Ulcer (DFU) is a critical risk factor for disability and mortality among diabetic patients, posing a significant public health challenge. Existing DFU datasets are limited in capturing the diversity and complexity of ulcer manifestations, preventing advancements in medical segmentation. This paper introduces SRRSH-DF, a novel DFU dataset that comprehensively covers characteristic symptoms and lesions across different foot regions. Expert participation ensures precise data annotations and professional evaluations. Building on this dataset, we develope UFOS-Net, a segmentation model incorporating the EMS Block to enhance the identification of small-scale masks and feature details. Additionally, we propose MODA, an improved data augmentation method tailored to the unique characteristics of DFU images. Experimental results demonstrate that UFOS-Net, when trained on the DFUC2022 dataset, ranks highly on the leaderboard. More importantly, training on SRRSH-DF achieved a Dice coefficient of 0.7745, validating the dataset's broad applicability and the model's effectiveness. We will provide the dataset to researchers who have a need for this study. This initiative aims to promote high-quality research and accelerate the advancement of DFU diagnostic tools, ultimately improving patient outcomes and reducing the burden of DFU in healthcare systems.