Artificial Intelligence Model for Assessing Early Childhood Development Through Children's Drawings in Kindergarten Levels 1–3 : Model Development And Validation Study
Article Sidebar
Main Article Content
Abstract
This study aimed (1) to develop an artificial intelligence (AI) model for classifying preschool children’s drawings from kindergarten levels 1–3 based on standardized tasks, and (2) to evaluate the model’s performance and feasibility for integration into a web-based early childhood development monitoring system. This applied research involved a purposive sample of 150 preschool children who completed five standardized drawing tasks, including straight lines, continuous curves, circles, squares, and triangles, following the official preschool teacher assessment manual. The research instruments included an EfficientNetV2-S-based AI model with data augmentation techniques, trained using CrossEntropyLoss and the Adam optimizer. The results indicated that the proposed model successfully classified children’s drawings into three developmental levels—advanced, intermediate, and beginner—with an overall accuracy of 0.61, precision of 0.58, recall of 0.61, and an F1-score of 0.59, while the validation loss ranged between 0.83 and 1.04 throughout the training process. Furthermore, the model demonstrated strong potential for deployment in web-based systems to support continuous developmental monitoring, reduce subjective assessment bias, and facilitate systematic planning of developmentally appropriate learning activities.
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
1. Any views and comments in the Journal of Social Innovation and Lifelong Learning are the authors’ views. The editorial staff have not to agree with those views and it is not considered as the editorial’s responsibility.
2. The responsibility of content and draft check of each article belongs to each author. In case, there is any lawsuit about copyright infringement. It is considered as the authors’ sole responsibility.
3. The article copyright belonging to the authors and The Far Eastern University are copyrighted legally. Republication must be received direct permission from the authors and The Far Eastern University in written form.
References
de Belen, R. A. J., Bednarz, T., Sowmya, A., & Del Favero, D. (2020). Computer vision in autism spectrum disorder research. Translational Psychiatry, 10(1), 333. https://doi.org/10.1038/s41398-020-01015-w
Chen, L., Xu, Y., & Zhang, H. (2021). Web-based deep learning system for automated childhood handwriting assessment: Accuracy evaluation and feasibility study. Computers in Human Behavior, 120, 106734. https://doi.org/10.1016/j.chb.2021.106734
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J., & Houlsby, N. (2021).
An image is worth 16×16 words: Transformers for image recognition at scale.
In Proceedings of the International Conference on Learning Representations (ICLR).
https://doi.org/10.48550/arXiv.2010.11929
Feng, X., Li, Y., & Zhou, W. (2024). Comprehensive assessment and early prediction of gross motor performance in toddlers with graph convolutional networks–based deep learning. JMIR Formative Research, 8, e51996. https://doi.org/10.2196/51996
He, K., Zhang, X., Ren, S., & Sun, J. (2016).Deep residual learning for image recognition.
In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 770–778). https://doi.org/10.1109/CVPR.2016.90
Liu, H., Wang, S., & Zhang, Y. (2023). AI-assisted research on intelligent evaluation system for preschool education. Education and Information Technologies, 28, 11589–11605. https://doi.org/10.1007/s10639-023-11875-6
Nguyen, T., & Chen, L. (2023). The key artificial intelligence technologies in early childhood education: Opportunities and challenges. Artificial Intelligence Review, 56, 12871–12899. https://doi.org/10.1007/s10462-023-10637-7
Puckett, M. B., & Black, J. K. (2020). The young child: Development from prebirth through age eight (7th ed.). Pearson Education.
Reinhart, L., Bischops, A. C., Kerth, J.-L., Hagemeister, M., Heinrichs, B., Eickhoff, S. B., Dukart, J., Konrad, K., Mayatepek, E., & Meissner, T. (2024). Artificial intelligence in child development monitoring: A systematic review on usage, outcomes and acceptance. Intelligence-Based Medicine, 9, 100134. https://doi.org/10.1016/j.ibmed.2024.100134
Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., & Chen, L.-C. (2018). MobileNetV2: Inverted residuals and linear bottlenecks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 4510–4520). https://doi.org/10.1109/CVPR.2018.00474
Shen, S., Zhou, Y., & Li, J. (2023). Multiview child motor development dataset for AI-driven analysis. GigaScience, 12(1), giad039. https://doi.org/10.1093/gigascience/giad039
Skinner, B. F. (1953). Science and human behavior. Free Press.
Trilling, B., & Fadel, C. (2009). 21st century skills: Learning for life in our times. Jossey-Bass.
Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press.
Wang, S., Liu, W., Wang, J., Yi, Y., & Xu, J. (2025). DrawAA-Net: An AI-supported evaluation tool for children’s drawings. In Proceedings of the International Conference on Artificial Intelligence in Education (AIED 2025) (pp. 309–322). Lecture Notes in Computer Science (LNAI, Vol. 15877). Springer. https://doi.org/10.1007/978-3-031-98414-3_22
Ye, J., Wang, C., & Wu, Y. (2018). Computer vision analysis captures atypical attention in toddlers with autism spectrum disorder. Journal of the American Academy of Child & Adolescent Psychiatry, 57(11), 896–904. https://doi.org/10.1016/j.jaac.2018.07.879
Zhang, Y., Wang, M., He, J., Li, N., Zhou, Y., & Huang, H. (2024).
AestheNet: Revolutionizing aesthetic perception diagnosis in education with hybrid deep nets. IEEE Transactions on Learning Technologies, 17, 2063–2075. https://doi.org/10.1109/TLT.2024.3405966
