CranioSynth: A Pipeline for Generating Synthetic Cranial Defects Through LLM-Assisted Parameterization and Anatomical Constraints
Resumo
A reconstrução craniana personalizada a partir de tomografias computadorizadas pode apoiar o planejamento de cranioplastias e a criação de próteses customizadas. Entretanto, bases de dados com defeitos cranianos variados, plausíveis e parametrizados ainda são limitadas, dificultando estudos computacionais e aplicações futuras de inteligência artificial em Bioinformática e Medicina. Este artigo apresenta o CranioSynth, um pipeline computacional para geração sintética de defeitos cranianos baseado em restrições anatômicas e Modelos de Linguagem de Larga Escala , ou Large Language Models (LLMs). O método processa exames DICOM, segmenta o osso, reconstrói o crânio em 3D e extrai informações geométricas. Com base nesses dados estruturados, um LLM sugere a localização e os parâmetros do defeito, enquanto a aplicação e a validação permanecem totalmente determinísticas.Referências
Fedorov, A. et al. (2012). 3d slicer as an image computing platform for the quantitative imaging network. Magnetic Resonance Imaging, 30(9):1323–1341.
Kropla, F. et al. (2023). Development of 3d printed patient-specific skull implants based on 3d surface scans. 3D Printing in Medicine, 9(19).
Li, J. et al. (2023). Towards clinical applicability and computational efficiency in automatic cranial implant design: An overview of the autoimplant 2021 cranial implant design challenge. Medical Image Analysis, 88:102865.
Liu, T. et al. (2024). 3d reconstruction of bone ct scan images based on deformable convex hull. Medical and Biological Engineering and Computing, 62:551–561.
Lorensen, W. E. and Cline, H. E. (1987). Marching cubes: A high resolution 3d surface construction algorithm. ACM SIGGRAPH Computer Graphics, 21(4):163–169.
Moor, M., Banerjee, O., Abad, Z. S. H., Krumholz, H. M., Leskovec, J., Topol, E. J., and Rajpurkar, P. (2023). Foundation models for generalist medical artificial intelligence. Nature, 616(7956):259–265.
Nimje, T. G. et al. (2024). Revolutionizing healthcare with 3d printing: Current applications and future prospects. In 2024 Parul International Conference on Engineering and Technology, pages 1–6.
Pinaya, W. H. L. et al. (2022). Brain Imaging Generation with Latent Diffusion Models, page 117–126. Springer Nature Switzerland.
Sharma, N. and Aggarwal, L. M. (2010). Automated medical image segmentation techniques. Journal of Medical Physics, 35(1):3–14.
van der Walt, S., Schönberger, J. L., Nunez-Iglesias, J., Boulogne, F., Warner, J. D., Yager, N., Gouillart, E., and Yu, T. (2014). scikit-image: Image processing in python. PeerJ, 2:e453.
Zhou, Q.-Y., Park, J., and Koltun, V. (2018). Open3d: A modern library for 3d data processing.
Kropla, F. et al. (2023). Development of 3d printed patient-specific skull implants based on 3d surface scans. 3D Printing in Medicine, 9(19).
Li, J. et al. (2023). Towards clinical applicability and computational efficiency in automatic cranial implant design: An overview of the autoimplant 2021 cranial implant design challenge. Medical Image Analysis, 88:102865.
Liu, T. et al. (2024). 3d reconstruction of bone ct scan images based on deformable convex hull. Medical and Biological Engineering and Computing, 62:551–561.
Lorensen, W. E. and Cline, H. E. (1987). Marching cubes: A high resolution 3d surface construction algorithm. ACM SIGGRAPH Computer Graphics, 21(4):163–169.
Moor, M., Banerjee, O., Abad, Z. S. H., Krumholz, H. M., Leskovec, J., Topol, E. J., and Rajpurkar, P. (2023). Foundation models for generalist medical artificial intelligence. Nature, 616(7956):259–265.
Nimje, T. G. et al. (2024). Revolutionizing healthcare with 3d printing: Current applications and future prospects. In 2024 Parul International Conference on Engineering and Technology, pages 1–6.
Pinaya, W. H. L. et al. (2022). Brain Imaging Generation with Latent Diffusion Models, page 117–126. Springer Nature Switzerland.
Sharma, N. and Aggarwal, L. M. (2010). Automated medical image segmentation techniques. Journal of Medical Physics, 35(1):3–14.
van der Walt, S., Schönberger, J. L., Nunez-Iglesias, J., Boulogne, F., Warner, J. D., Yager, N., Gouillart, E., and Yu, T. (2014). scikit-image: Image processing in python. PeerJ, 2:e453.
Zhou, Q.-Y., Park, J., and Koltun, V. (2018). Open3d: A modern library for 3d data processing.
Publicado
19/10/2026
Como Citar
PEREIRA, Marcus V. C.; BEZERRA, Marcus M.; SANTOS, Danilo F. S..
CranioSynth: A Pipeline for Generating Synthetic Cranial Defects Through LLM-Assisted Parameterization and Anatomical Constraints. In: ENCONTRO NACIONAL DE INTELIGÊNCIA ARTIFICIAL E COMPUTACIONAL (ENIAC), 23. , 2026, Cuiabá/MT.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2026
.
p. 2871-2882.
ISSN 2763-9061.
DOI: https://doi.org/10.5753/eniac.2026.30782.
