Diagnóstico Automático de Fraturas do Escafoide em Radiografias do Punho Usando Arquiteturas de Redes Neurais Convolucionais

• Luiz F. A. Guerra UNDB
• Carine de O. Vieira UNDB
• Augusto C. R. Xavier UNDB
• Giovanni L. F. da Silva UNDB
• João O. B. Diniz IFMA

DOI: https://doi.org/10.5753/sbcas.2026.21719

Resumo

As fraturas do escafoide representam um desafio persistente na radiografia do punho, especialmente em casos não deslocados e em estágios iniciais do trauma. Diante disso, este trabalho propõe um método baseado em Redes Neurais Convolucionais para detecção automática e diagnóstico de fraturas em radiografias do punho. O método proposto emprega Faster R-CNN com backbone ResNet-50 para detecção do escafoide e InceptionResNetV2 para classificação. Resultados preliminares demonstraram bom desempenho, com IoU de 93,1% na detecção e acurácia de 98,22% no diagnóstico. Os achados indicam potencial da abordagem para apoio ao diagnóstico radiográfico.

Referências

Alshamrani, H. A. (2023). Accuracy of the radiological protocols in detecting scaphoid fractures, a retrospective study. Signa Vitae, 19(4).

Bützow, A., Anttila, T., Haapamäki, V., and Ryhänen, J. (2025). A novel segmentation-based deep learning model for enhanced scaphoid fracture detection. European Journal of Radiology, page 112309.

Gomaa, A., Minematsu, T., Abdelwahab, M. M., Abo-Zahhad, M., and Taniguchi, R.-i. (2022). Faster cnn-based vehicle detection and counting strategy for fixed camera scenes. Multimedia Tools and Applications, 81(18):25443–25471.

Hendrix, N., Scholten, E., Vernhout, B., Bruijnen, S., Maresch, B., de Jong, M., Diepstraten, S., Bollen, S., Schalekamp, S., de Rooij, M., et al. (2021).

Development and validation of a convolutional neural network for automated detection of scaphoid fractures on conventional radiographs. Radiology: Artificial Intelligence, 3(4):e200260.

Kraus, M., Anteby, R., Konen, E., Eshed, I., and Klang, E. (2024). Artificial intelligence for x-ray scaphoid fracture detection: a systematic review and diagnostic test accuracy meta-analysis. European radiology, 34(7):4341–4351.

Lee, K.-C., Choi, I. C., Kang, C. H., Ahn, K.-S., Yoon, H., Lee, J.-J., Kim, B. H., and Shim, E. (2023). Clinical validation of an artificial intelligence model for detecting distal radius, ulnar styloid, and scaphoid fractures on conventional wrist radiographs. Diagnostics, 13(9):1657.

Mallee, W. H., Mellema, J. J., Guitton, T. G., Goslings, J. C., Ring, D., Doornberg, J. N., and of Variation Group, S. (2016). 6-week radiographs unsuitable for diagnosis of suspected scaphoid fractures. Archives of orthopaedic and trauma surgery, 136(6):771–778.

Oeding, J. F., Kunze, K. N., Messer, C. J., Pareek, A., Fufa, D. T., Pulos, N., and Rhee, P. C. (2024). Diagnostic performance of artificial intelligence for detection of scaphoid and distal radius fractures: a systematic review. The Journal of Hand Surgery, 49(5):411–422.

Rodrigo, M. S. et al. (2024). Bone fracture multi-region x-ray data. Kaggle dataset. Accessed: 28 Sep 2025.

Roh, M.-C. and Lee, J.-y. (2017). Refining faster-rcnn for accurate object detection. In 2017 fifteenth IAPR international conference on machine vision applications (MVA), pages 514–517. IEEE.

Scaphoids (2023). Scaphoid-scanner dataset. [link]. visited on 2025-10-29.

Sharma, S. (2023). Artificial intelligence for fracture diagnosis in orthopedic x-rays: current developments and future potential. SICOT-J, 9:21.

Yang, T.-H., Horng, M.-H., Li, R.-S., and Sun, Y.-N. (2022). Scaphoid fracture detection by using convolutional neural network. Diagnostics, 12(4):895.

Yang, T.-H., Sun, Y.-N., Li, R.-S., and Horng, M.-H. (2024). The detection and classification of scaphoid fractures in radiograph by using a convolutional neural network. Diagnostics, 14(21):2425.

Żyluk, A. (2023). Diagnostics and management of acute scaphoid fractures: an update. Pomeranian Journal of Life Sciences, 69(3).