Autism Spectrum Disorder (ASD) is a neurodevelopmental condition characterized by challenges in social interaction, communication, and behavioral patterns. Early diagnosis plays a crucial role in providing timely intervention and improving developmental outcomes. This project proposes a software-based intelligent system for early autism detection using facial image analysis and deep learning techniques. The system utilizes a publicly available facial image dataset containing labeled images of children with and without ASD. Image preprocessing techniques such as resizing, normalization, and face region extraction are applied to improve model performance. A Convolutional Neural Network (CNN) model, implemented using transfer learning with MobileNetV2 architecture, is employed for feature extraction and classification. The model is trained to distinguish between ASD and non-ASD cases based on visual behavioral cues present in facial patterns. The backend is developed using Python and Flask, while the frontend interface is built using React.js for image upload and result visualization. The trained model outputs a probability score indicating the likelihood of ASD, along with classification results. System performance is evaluated using metrics such as accuracy, precision, recall, F1-score, and confusion matrix analysis. The proposed solution provides a cost-effective, non-invasive, and scalable screening tool that can assist healthcare professionals and parents in early-stage autism risk assessment. This system demonstrates the potential of deep learning in supporting clinical decision-making and advancing intelligent healthcare applications modules.

The proposed system presents an efficient AI-based solution for the early detection of Autism Spectrum Disorder (ASD) using facial image analysis. By leveraging the MobileNetV2 deep learning model with transfer learning, the system is capable of accurately classifying ASD and Non-ASD cases. The integration of image preprocessing, feature extraction, and classification ensures reliable and consistent performance. Furthermore, the implementation of a web-based interface enables real-time prediction and user-friendly interaction. The system also provides a probability-based risk assessment, making the output more interpretable and useful for early screening. Overall, the proposed solution is cost-effective, scalable, and contributes to the advancement of intelligent healthcare systems.

[1]D. Thabtah, “Machine Learning in Autism Spectrum Disorder Detection,” International Journal of Intelligent Systems and Applications, vol. 11, no. 3, pp. 1–10, 2019. [2] M. Sandler et al., “MobileNetV2: Inverted Residuals and Linear Bottlenecks,” in Proc. IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2018, pp. 4510–4520. [3]K. He et al., “Deep Residual Learning for Image Recognition,” in Proc. IEEE Conf. on ComputerVision and Pattern Recognition (CVPR), 2016, pp. 770–778. [4]A. Krizhevsky, I. Sutskever, and G. Hinton, “ImageNet Classification with Deep Convolutional Neural Networks,” in Advances in Neural Information Processing Systems (NIPS), 2012, pp. 1097–1105. [5]R. Singh and S. Kumar, “Image-Based Autism Detection Using Convolutional Neural Networks,” in Proc. International Conference on Intelligent Computing and Control Systems (ICICCS), India, 2021, pp. 1023–1027.

© 2025 International Journal of Engineering and Techniques (IJET).

Digital Object Identifier (DOI)

IJET assigns a unique DOI to every accepted article via Zenodo for persistent linking and citation. Your article’s DOI: https://doi.org/{{doi}}

Open DOI About Zenodo DOI

Close