Scalable and Efficient Deep Learning for Diabetic Retinopathy Classification on ARM-Based Architectures

• Thiago da Silva Araújo UFRGS
• Philippe O. A. Navaux UFRGS

DOI: https://doi.org/10.5753/sbcas_estendido.2026.21203

Resumo

Diabetic retinopathy (DR) is a leading cause of preventable blindness and a public health problem in Brazil, where access to ophthalmological specialists is often limited. Automated DR classification using deep learning (DL) can support screening programs, but deployment requires accurate, efficient, and scalable models. This work evaluates model selection and distributed scalability of convolutional neural networks (CNNs) for binary DR classification on ARM-based systems, using datasets from Federal University of São Paulo (UNIFESP), Hospital de Clínicas de Porto Alegre (HCPA), and TeleOftalmo (TelessaúdeRS-UFRGS). Among 38 evaluated CNN architectures, MobileNet performed best, consuming 77% less energy, training 83% faster, and producing an 85% smaller model than InceptionV3, while improving AUC by 3%. A preliminary distributed scalability evaluation across two NVIDIA Grace Superchips using data parallelism achieved a 1.93× speedup and 96.5% scaling efficiency. These results indicate that ARM-based distributed training is a feasible, energy-efficient option for scalable DR screening in Brazilian healthcare.

Referências

Abushawish, I. Y., Modak, S., Abdel-Raheem, E., Mahmoud, S. A., and Hussain, A. J. (2024). Deep learning in automatic diabetic retinopathy detection and grading systems: a comprehensive survey and comparison of methods. IEEE Access, 12:84785–84802.

Arora, L., Singh, S. K., Kumar, S., Gupta, H., Alhalabi, W., Arya, V., Bansal, S., Chui, K. T., and Gupta, B. B. (2024). Ensemble deep learning and efficientnet for accurate diagnosis of diabetic retinopathy. Scientific Reports, 14(1):30554.

Banchelli, F., Vinyals-Ylla-Catala, J., Pocurull, J., Clascà, M., Peiro, K., Spiga, F., Garcia-Gasulla, M., and Mantovani, F. (2024). Nvidia grace superchip early evaluation for hpc applications. In Proceedings of the International Conference on High Performance Computing in Asia-Pacific Region Workshops, pages 45–54.

Dos Reis, M. A., Künas, C. A., et al. (2024). Advancing healthcare with artificial intelligence: diagnostic accuracy of machine learning algorithm in diagnosis of diabetic retinopathy in the brazilian population. Diabetology & Metabolic Syndrome, 16(1):209.

Flaxman, S. R., Bourne, R. R., Resnikoff, S., et al. (2017). Global causes of blindness and distance vision impairment 1990–2020: a systematic review and meta-analysis. The Lancet Global Health, 5(12):e1221–e1234.

Kallel, F. and Echtioui, A. (2024). Retinal fundus image classification for diabetic retinopathy using transfer learning technique. Signal, image and video processing, 18(2):1143–1153.

Lutz de Araujo, A., Moreira, T. d. C., Varvaki Rados, D. R., Gross, P. B., Molina-Bastos, C. G., Katz, N., Hauser, L., Souza da Silva, R., Gadenz, S. D., Dal Moro, R. G., et al. (2020). The use of telemedicine to support brazilian primary care physicians in managing eye conditions: The teleoftalmo project. PloS one, 15(4):e0231034.

Massuda, A., Hone, T., et al. (2018). The brazilian health system at crossroads: progress, crisis and resilience. BMJ Global Health, 3(4):e000829.

Nagpal, D., Panda, S. N., Malarvel, M., Pattanaik, P. A., and Khan, M. Z. (2022). A review of diabetic retinopathy: Datasets, approaches, evaluation metrics and future trends. Journal of King Saud University-Computer and Information Sciences, 34(9):7138–7152.

Nakayama, L. F., Goncalves, M., Zago Ribeiro, L., Santos, H., Ferraz, D., Malerbi, F., Celi, L. A., and Regatieri, C. (2023). A brazilian multilabel ophthalmological dataset (brset). PhysioNet DOI: 10.13026/xcxw-8198, 13026.

Patil, M. S. and Chickerur, S. (2023). Study of data and model parallelism in distributed deep learning for diabetic retinopathy classification. Procedia Computer Science, 218:2253–2263.

Ruhela, A., Cazes, J., McCalpin, J., del Castillo-Negrete, C., Li, J., Liu, H., Chen, H., Lu, C.-Y., Milfeld, K., Zhang, W., et al. (2024). Performance analysis of scientific applications on an nvidia grace system. In SC24-W: Workshops of the International Conference for High Performance Computing, Networking, Storage and Analysis, pages 558–566. IEEE.

Saproo, D., Mahajan, A. N., and Narwal, S. (2024). Deep learning based binary classification of diabetic retinopathy images using transfer learning approach. Journal of Diabetes & Metabolic Disorders, 23(2):2289–2314.

Shanthala, K. and Kundur, N. C. (2025). Dr-efficientnet-l: A distributed deep learning architecture for efficient detection and grading of diabetic retinopathy. Engineering, Technology & Applied Science Research, 15(5):28362–28367.