Comparing Video-Based Representations in Classification of General Movements of Newborns

• Beatriz Emily Silva Aguiar UFAM
• Eulanda M. Santos UFAM
• Rafael Giusti UFAM
• Ayrles Mendonça UFAM

DOI: https://doi.org/10.5753/sbcas.2025.7732

Abstract

General Movements Assessment (GMA) is a clinical method for early detection of neuromotor disorders in infants. In this study we compare video-based representations consolidated in the literature applied to classify Writhing Movements (WMs): RGB videos, optical flow, keypoints, and motion histograms. In our experiments conducted using a proprietary dataset, the RGB-based model achieved the highest individual accuracy (0.83), while the model trained using histograms obtained the best F1-Score (0.83) and efficiency. Finally, the fusion of the models via ensemble increased accuracy to 0.87. These results indicate that integrating multiple representations can enhance automated GMA analysis and contribute to the early diagnosis of neuromotor disorders.

References

Camargos, F. O. and Almeida, J. G. A. d. (2019). Paralisia cerebral: revisão e considerações atuais. Acta Fisiátrica, 26(3):144–152.

Chopard, D., Laguna, S., Chin-Cheong, K., Dietz, A., Badura, A., Wellmann, S., and Vogt, J. E. (2024). Automatic classification of general movements in newborns. In Findings of the AHLI Machine Learning for Health (ML4H) Symposium, Vancouver, Canada.

Doroniewicz, I., Ledwoń, D. J., Affanasowicz, A., Kieszczyńska, C., Latos, D., Matyja, M., Mitas, A. W., and Myśliwiec, A. (2020). Writhing movement detection in newborns on the second and third day of life using pose-based feature machine learning classification. Sensors, 20(21):5986.

Ferrari, F. and Cioni, G. (2016). Fidgety movements—tiny in appearance, but huge in impact. Journal of Pediatrics, 92:S64–S70.

Hadders-Algra, M. (2004). General movements: A window for early identification of children at high risk for developmental disorders. The Journal of Pediatrics, 145(2 Suppl):S12–S18.

Hashimoto, Y., Ishikawa, K., et al. (2022). Automated classification of general movements in infants using two-stream spatiotemporal fusion network. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 30:456–465.

Hesse, N., Pujades, S., Black, M. J., Arens, M., Hofmann, U. G., and Schroeder, A. S. (2021). Learning and tracking the 3d body shape of freely moving infants from rgb-d sequences. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 14703–14713.

Li, P., Xu, Y., Wei, Y., and Yang, Y. (2022). Self-correction for human parsing. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(6):3260–3271.

McCay, K. D., Hu, P., Shum, H. P. H., Ho, E. S. L., Woo, W. L., Marcroft, C., Embleton, N. D., and Munteanu, A. (2022). A pose-based feature fusion and classification framework for the early prediction of cerebral palsy in infants. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 30:8–19.

Moryossef, A. et al. (2023). Poseformat: Library for viewing, augmenting, and handling pose files. arXiv preprint arXiv:2310.09066.

Orlandi, S., Cinque, L., Ferrante, G., Sgandurra, G., and Cioni, G. (2018). Detection of atypical and typical infant movements using computer-based video analysis. Journal of Medical Imaging, 5(2):024001.

Palheta, M., Santos, G., Mendonça, A., Gonçalves, P., Albuquerque, R., Souto, E., and Santos, E. (2023). Fusão de dados de vídeos rgb e pontos-chaves para classificação de movimentos gerais de bebês. In Anais do XXIII Simpósio Brasileiro de Computação Aplicada à Saúde, pages 384–394, Porto Alegre, RS, Brasil. SBC.

Prechtl, H. F. (1997). General movements: A window for early detection of developmental disorders. Pediatrics, 99:92–101.

Raghuram, K., Orlandi, S., Church, P., Luther, M., Kiss, A., and Shah, V. (2022). Automated movement analysis to predict cerebral palsy in very preterm infants: An ambispective cohort study. Children, 9(6):843–851.

Raghuram, K., Orlandi, S., Shah, V., Chau, T., Luther, M., Banihani, R., and Church, P. (2019). Automated movement analysis to predict motor impairment in preterm infants: A retrospective study. Journal of Perinatology, 39(10):1362–1369.

Silva, N., Zhang, D., Kulvicius, T., Gail, A., Barreiros, C., Lindstaedt, S., Kraft, M., Bölte, S., Poustka, L., Nielsen-Saines, K., Wörgötter, F., Einspieler, C., and Marschik, P. (2021). The future of general movement assessment: The role of computer vision and machine learning - a scoping review. Research in Developmental Disabilities, 110.

Solovyova, S. et al. (2020). Eye-tracking metrics in children with autism spectrum disorder. arXiv preprint arXiv:2008.09670.

Spittle, A. J. and Doyle, L. W. (2018). Identification of neurodevelopmental impairments in preterm infants. Journal of Pediatrics, 191:20–28.

Warrington, H. et al. (2021). A systematic review of automated methods for general movements assessment. Developmental Medicine & Child Neurology, 63:745–756.

Washington, P. et al. (2021). Computer-based analysis of eye tracking to detect early signs of autism spectrum disorder. Journal of Autism and Developmental Disorders, 51(5):1435–1450.

Zhu, M., Men, Q., Ho, E. S. L., Leung, H., and Shum, H. P. H. (2021). Interpreting deep learning based cerebral palsy prediction with channel attention. In 2021 IEEE EMBS Inter-national Conference on Biomedical and Health Informatics (BHI), pages 1–4. IEEE.