Automated Human Figure Drawing Test Assessment System through a Computational Vision Model
Keywords:
Children psychology, Computer vision, Human figure drawings, Machine learning, TechnologyAbstract
Cognitive disorders in childhood, particularly intellectual disability in Mexico and other countries, represent a significant public health challenge. These are often not adequately detected in early stages, limiting timely interventions crucial for the development and well-being of children. The projective technique of Human Figure Drawing (HFD), used by psychologists for the early detection of cognitive problems in children, is limited by a time-consuming process and a subjective manual analysis. In response to these challenges, this paper details the development of an automated system for the evaluation of HFD in children aged 5 to 12 years in the Metropolitan Area of Nuevo León, Mexico, with the aim of improving the early detection of cognitive disorders. The hypothesis proposed is that the developed tool will enable the identification of HFD graphic indicators with an accuracy of over 90%, reducing the evaluation time from 10 minutes to less than 30 seconds. The methodology involves the collection of over 1,000 drawings from the application of HFD in the target population, followed by their processing and labeling to convert them into training data for a machine learning model. As part of the results, a computer vision model was developed, capable of detecting up to 24 graphic indicators in a human figure drawing in less than 1 second. The main findings of the study reveal a reduction in the evaluation task time of an HFD by 99.83%, and a mean Average Precision (mAP50) of 92.8, a precision (P) of 0.912 and a recall (R) of 0.902, highlighting the model’s effectiveness in correctly identifying the elements in the drawings. The complete system offers a web application dedicated to evaluating the drawings, thus facilitating the user’s interaction with the proposed model. The system integrates technologies such as Python, PyTorch, and YOLO for the processing and analysis of the drawings, and Vue.js, C#, and SQL Server for the web application development. This technological advance presents significant potential applications in hospitals, clinics, mental health centers, and educational settings. It represents a valuable tool for diagnosing and treating cognitive disorders in children, improving not only the efficiency and objectivity in HFD evaluation but also opening new possibilities for addressing challenges in children’s mental health.
References
[1] Gaytán, E., Rosales, M, Reyes, H., Díaz-Barriga, F., & Calderón, J. (2015). Prevalencia de dificultades emocionales, conductuales y cognitivas en niños de escenarios urbanos con diferente grado de marginación. Revista de psicología y ciencias del comportamiento de la Unidad Académica de Ciencias Jurídicas y Sociales, 6(1), 57-74. http://www.scielo.org.mx/scielo.php?script=sci_arttext&pid=S2007-18332015000100004&lng=es&tlng=es.
[2] Delahunty, L., O’Hare, A., Marryat, L., Stewart, T. M., McKenzie, K., Murray, G., & Kaza, N. (2022). Short Report: Exploring the extent to which Intellectual Disability is undiagnosed within children attending developmental pediatric clinics. Research in Developmental Disabilities, 131, 104359. https://doi.org/10.1016/j.ridd.2022.104359
[3] Chollat, C., Joly, A., Houivet, E., Bénichou, J., & Marret, S. (2019). School-age human figure drawings by very preterm infants: Validity of the Draw-a-Man test to detect behavioral and cognitive disorders. Archives de Pédiatrie, 26(4), 220-225. https://doi.org/10.1016/j.arcped.2019.02.015.
[4] Dağlioğlu, H. E., Deniz, Ü., & Kan, A. (2010). A study on the emotional indicators in 5-6 year-old girls’ and boys’ human figure drawings. Procedia - Social and Behavioral Sciences, 2(2), 1503-1510. https://doi.org/10.1016/j.sbspro.2010.03.226
[5] Quiroga, A., Fuentes, M., Saucedo, E., Briones, E. N., Lozano, A. P., Montegudo, D., & González, N. (2022). Sistematización de la Evaluación Psicológica de Recursos Personales a través del Dibujo de la Persona. Universidad de Monterrey, Vicerrectoría de Ciencias de la Salud, Escuela de Psicología. Programa de Evaluación Final del Programa Académico de Licenciatura en Psicología/Licenciatura en Psicología Clínica. San Pedro Garza García, Nuevo León.
[6] Raschka, S., & Mirjalili, V. (2017). Python Machine Learning (2nd ed.). Packt Publishing.
[7] Taye, M. M. (2023). Theoretical understanding of convolutional neural network: Concepts, architectures, applications, future directions. Computation, 11(3), 52. https://doi.org/10.3390/computation11030052
[8] Luxton, D. D. (2014). Artificial intelligence in psychological practice: Current and future applications and implications. Professional Psychology: Research and Practice, 45(5), 332–339. https://doi.org/10.1037/a0034559
[9] Hamet, P., & Tremblay, J. (2017). Artificial intelligence in medicine. Metabolism, 69(Supplement), 36-40. https://doi.org/10.1016/j.metabol.2017.01.011
[10] Faust, O., Hagiwara, Y., Hong, T. J., Lih, O. S., & Acharya, U. R. (2018). Deep learning for healthcare applications based on physiological signals: A review. Computer Methods and Programs in Biomedicine, 161, 1-13. https://doi.org/10.1016/j.cmpb.2018.04.005
[11] Kissos, L., Goldner, L., Butman, M., Eliyahu, N., & Lev-Wiesel, R. (2020). Can artificial intelligence achieve human-level performance? A pilot study of childhood sexual abuse detection in self-figure drawings. Child Abuse & Neglect, 109, 104755. https://doi.org/10.1016/j.chiabu.2020.104755
[12] Zeeshan, M.O., Siddiqi, I., & Moetesum, M. (2021). Two-Step Fine-Tuned Convolutional Neural Networks for Multi-label Classification of Children’s Drawings. In J. Lladós, D. Lopresti, & S. Uchida (Eds.), Document Analysis and Recognition – ICDAR 2021, 12822, 321-334. Springer, Cham. https://doi.org/10.1007/978-3-030-86331-9_21
[13] Romero, V. (2019). How Artificial Intelligence can detect emotions in children’s drawings. https://medium.datadriveninvestor.com/how-artificial-intelligence-can-detect-emotions-in-childrens-drawings-4359cf51ab3d
[14] Taeuck, L. (2022). Understanding Children's Psychology with Drawings Drawn by Children Using AI. https://www.24-7pressrelease.com/press-release/497168/korea-contact-free-insighter-understanding-childrens-psychology-with-drawings-drawn-by-children-using-ai
[15] Moetesum, M., Aslam, T., Saeed, H., Siddiqi, I., & Masroor, U. (2017). Sketch-based Facial Expression Recognition for Human Figure Drawing Psychological Test. In 2017 International Conference on Frontiers of Information Technology (FIT), 258-263. Islamabad, Pakistan. https://doi.org/10.1109/FIT.2017.00053
[16] Widiyanto, S., & Abuhasan, J. W. (2020). Implementation The Convolutional Neural Network Method For Classification The Draw-A-Person Test. In 2020 Fifth International Conference on Informatics and Computing (ICIC), 1-6. Gorontalo, Indonesia. https://doi.org/10.1109/ICIC50835.2020.9288651
[17] Schwaber, K. & Sutherland, J. (2020). The 2020 Scrum Guide. https://scrumguides.org/scrum-guide.html
[18] Beck, K., & Andres, C. (2004). Extreme Programming Explained: Embrace Change (2nd ed.). Addison-Wesley Professional.
[19] Terra, J. (2023). Keras vs Tensorflow vs Pytorch: Key Differences Among Deep Learning. https://www.simplilearn.com/keras-vs-tensorflow-vs-pytorch-article
[20] Boesch, G. (2023). Object Detection in 2023: The Definitive Guide. https://viso.ai/deep-learning/object-detection/
[21] Elyan, E., Jamieson, L. and Ali-Gombe, A. (2020). Deep learning for symbols detection and classification in engineering drawings. Neural networks, 129, 91-102. https://doi.org/10.1016/j.neunet.2020.05.025
[22] Zhao, Y., Deng, X., & Huahui, L. (2020). A YOLO-Based Method to Recognize Structural Components from 2D Drawings. In Construction Research Congress 2020: Computer Application, 753-762. https://doi.org/10.1061/9780784482865.080
[23] Schcolnik-Elias, A., Martinez, S., Luna-Taylor, J., & Castro, I. (2023). Detección de armas tipo pistola mediante el uso de redes convolucionales con una arquitectura tipo YOLO y estereoscopía. Pädi Boletín Científico de Ciencias Básicas e Ingenierías del ICBI, 11(2), 196-204. https://doi.org/10.29057/icbi.v11iEspecial2.10727
[24] King, R. (2023). Brief summary of YOLOv8 model structure. https://github.com/ultralytics/ultralytics/issues/189
[25] Casas, E., Ramos, L., Bendek, E., & Rivas-Echeverría, F. (2023). Assessing the Effectiveness of YOLO Architectures for Smoke and Wildfire Detection. IEEE Access, 11, 96554-96583. https://doi.org/10.1109/ACCESS.2023.3312217
[26] Shorten, C. & Khoshgoftaar, T. M. (2019). A survey on Image Data Augmentation for Deep Learning. Journal of Big Data, 6(60). https://doi.org/10.1186/s40537-019-0197-0
[27] Kearns, M. & Nenkova, A. (2019). Training, Validation and Test sets. University of Pennsylvania. https://www.cis.upenn.edu/~mkearns/teaching/ScienceDataEthics/files/lecture /l5.pdf
