Classification of Diabetic Wounds Using Transfer Learning Model: EfficientNetB1 and ResNet50

George Morris William Tangka; Andreas Rantepadang; Green Arther Sandag; Rolly Junius Lontaan; Wilsen Grivin Mokodaser

doi:10.31154/cogito.v11i1.1001.207-217

Authors

George Morris William Tangka Universitas Klabat
Andreas Rantepadang Universitas Klabat
Green Arther Sandag Universitas Klabat
Rolly Junius Lontaan Universitas Klabat
Wilsen Grivin Mokodaser Universitas Klabat

DOI:

https://doi.org/10.31154/cogito.v11i1.1001.207-217

Keywords:

Diabetic Wound, Transfer Learning, Deep Learning, ResNet50, EfficientNetB1

Abstract

Diabetic foot ulcers are a major complication of diabetes, and their early detection remains difficult in routine practice. We address this gap by evaluating transfer learning models—EfficientNet-B1 and ResNet-50—for automated diabetic-wound classification. Using a two-class image dataset (“Diabetic Wounds” vs. “Normal”), we fine-tuned both backbones with two optimizers (SGD, Adamax). Models were trained for 50 epochs (batch size 16) with standard data augmentation to improve generalization. Performance was evaluated by classification accuracy. EfficientNet-B1 with SGD achieved the best test accuracy (99.48%), outperforming EfficientNet-B1 with Adamax (98.86%), ResNet-50 with SGD (99.22%), and ResNet-50 with Adamax (97.66%). These results indicate that transfer learning—particularly EfficientNet-based architectures optimized with SGD—can provide highly accurate, automated screening of diabetic wounds. The approach shows promise for integration into clinical decision-support systems to assist timely triage and management, and motivates future work on multi-center, patient-level validation and evaluation across diverse skin tones and imaging conditions.

References

World Health Organization, “Diabetes,” 2023. [Online]. Available: https://www.who.int/news-room/fact-sheets/detail/diabetes.

N. Singh, D. G. Armstrong, and B. A. Lipsky, “Preventing Foot Ulcers in Patients with Diabetes,” JAMA, vol. 293, no. 2, pp. 217–228, Jan. 2005, doi: 10.1001/jama.293.2.217.

A. Rantepadang, “Perbandingan Kualitas Hidup Pasien Gagal Ginjal Kronik antara Komorbid Faktor Diabetes Mellitus dan Hipertensi pada Pasien yang Menjalani Hemodialisa,” Nutrix Journal, vol. 5, no. 2, pp. 1–7, Oct. 2021. [Online]. Available: https://ejournal.unklab.ac.id/index.php/nutrix/article/view/575.

H. C. Shin et al., “Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning,” IEEE Trans. Med. Imaging, vol. 35, no. 5, pp. 1285–1298, May 2016, doi: 10.1109/TMI.2016.2528162.

S. J. Pan and Q. Yang, “A Survey on Transfer Learning,” IEEE Trans. Knowl. Data Eng., vol. 22, no. 10, pp. 1345–1359, Oct. 2010, doi: 10.1109/TKDE.2009.191.

G. A. Sandag, E. Djamal, G. M. W. Tangka, dan S. W. Taju, "MRI Image Analysis for Alzheimer's Disease Detection Using Transfer Learning: VGGNet vs. EfficientNet," COGITO Smart Journal, vol. 10, no. 2, hlm. 580-592, Des. 2024.

C. Wang, Z. Yu, Z. Long, and H. Zhao, “A Few-shot Diabetes Foot Ulcer Image Classification Method Based on Deep ResNet and Transfer Learning,” J. King Saud Univ. – Comput. Inf. Sci., vol. 37, no. 2, pp. 187–198, 2024.

Z. Liu, J. John, and E. Agu, “Diabetic Foot Ulcer Ischemia and Infection Classification Using EfficientNet Deep Learning Models,” IEEE Open J. Eng. Med. Biol., vol. 3, pp. 179–189, 2022, doi: 10.1109/OJEMB.2022.3147496.

S. Ullah, A. Javed, M. Aljasem, and A. K. J. Saudagar, “Eff-ReLU-Net: A Deep Learning Framework for Multiclass Wound Classification,” BMC Medical Imaging, vol. 25, no. 1, 2025, Art. no. 45, doi: 10.1186/s12880-025-01785-z.

S. Debnath, A. Khurana, M. Senbagavalli, P. K. Choudhary, and R. D. Dey, “Sustainable AI for Diabetic Foot Ulcer Detection: A Deep Learning Approach for Early Diagnosis,” Discover Applied Sciences, vol. 5, no. 2, 2025, Art. no. 148, doi: 10.1007/s44202-025-00363-1.

M. Tan and Q. V. Le, “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,” in Proc. 36th Int. Conf. on Machine Learning, Long Beach, CA, USA, May 2019, pp. 6105–6114. [Online]. Available: http://proceedings.mlr.press/v97/tan19a.html

Alzubaidi, L., Fadhel, M. A., Oleiwi, S. R., Al-Shamma, O., & Zhang, J. (2020). DFU_QUTNet: diabetic foot ulcer classification using novel deep convolutional neural network. Multimedia Tools and Applications, 79(21), 15655-15677.

L. Alzubaidi et al., "Robust application of new deep learning tools: an experimental study in medical imaging," Multimedia Tools Appl., vol. 80, pp. 26365–26393, 2021.

L. Alzubaidi et al., "Towards a better understanding of transfer learning for medical imaging: a case study," Appl. Sci., vol. 10, no. 13, p. 4523, 2020.

S. Debnath et al., "Sustainable AI for diabetic foot ulcer detection: a deep learning approach for early diagnosis," Discover Appl. Sci., vol. 7, no. 1012, 2025.

N. Almufadi and H. F. Alhasson, "Classification of diabetic foot ulcers from images using machine learning approach," Diagnostics, vol. 14, no. 16, Art. no. 1807, Aug. 2024, doi: 10.3390/diagnostics14161807.

B. Sistaninejhad, H. Rasi, and P. Nayeri, “A review paper about deep learning for medical image analysis,” Computational and Mathematical Methods in Medicine, vol. 2023, Article ID 7091301, pp. 1–10, May 2023. doi: 10.1155/2023/7091301.

I. D. Mienye et al., "Deep Convolutional Neural Networks: A Comprehensive Review," Preprints.org, preprint, Aug. 19, 2024, doi: 10.20944/preprints202408.1288.v1.

M. Tan and Q. V. Le, "EfficientNet: Rethinking model scaling for convolutional neural networks," in Proceedings of the 36th International Conference on Machine Learning, Long Beach, CA, USA, 2019, pp. 6105–6114.

L. Zhang, Y. Bian, P. Jiang, and F. Zhang, "A Transfer Residual Neural Network Based on ResNet-50 for Detection of Steel Surface Defects," Appl. Sci., vol. 13, no. 9, p. 5260, Apr. 2023, doi: 10.3390/app13095260.

M. Feng, Y. Cai, and S. Yan, "Enhanced ResNet50 for Diabetic Retinopathy Classification: External Attention and Modified Residual Branch," Mathematics, vol. 13, no. 10, Art. no. 1557, May 2025, doi: 10.3390/math13101557.