Computer Vision Model for the Pre-diagnosis of Autism from Facial Images

Arellano F.; Huaman A.; Molina P.; Ticona W. · Lecture Notes in Networks and Systems (2025) · DOI: 10.1007/978-3-031-96798-6_38

Resumen

Autism spectrum disorder (ASD) is a neurodevelopmental condition that significantly affects communication and social interaction. Early detection is crucial to provide timely interventions to improve the quality of life of affected individuals. However, diagnosis of ASD is challenging due to the lack of specific medical tests, which often result in late diagnosis. In this context, the aim of this study is to implement a computational vision model for the pre-diagnosis of autism from facial images. The applied methodology was based on 4 phases: Data collection, Preprocessing, Implementation (VGG-19, Xception and Vision Transformer) and evaluation (Accuracy, Recall, Precision and F1-Score). The results demonstrate that the proposed Xception model outperforms the others, achieving an accuracy of 73% and an F1 score of 76%, which indicates its effectiveness in accurately identifying ASD cases. This approach not only reduces the time and resources needed for a clinical diagnosis but also provides a basis for future research on noninvasive and accessible diagnostic tools for autism. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

Instituciones: Universidad Tecnológica del Perú