Autism Spectrum Disorder Prediction from Facial Images Using Fine-Tuned Efficient Net B0–B7 ‎Architectures

  • Authors

    • V. Krishnamoorthy Professor, Dean R&D, Department of ECE, St.Michael College of Engineering and Technology, Kalayarkoil, Sivagangai District, India
    • T. Veeramani Professor, Department of Artificial Intelligence and Data Science, Panimalar Engineering College
    • M. Indirani Assistant Professor, Department of Computer Science and Business Systems, Vel Tech Multi Tech ‎ Dr.Rangarajan Dr.Sakunthala Engineering College, Avadi, Chennai
    • B. R. Sathishkumar Associate Professor, Department of Electronics and Communication Engineering, Sri Ramakrishna Engineering ‎College, Coimbatore - 641022, Tamil Nadu, India
    • G. Sasi Professor, Department of Electronics and Communication Engineering, Chettinad Institute of Technology, Chettinad Academy of‎ Research and Education, Manamai off Campus, ECR, Chengalpattu- 603 102, Tamil Nadu, India
    • K. Selvakumarasamy Professor, Department of Electronics and Communication Engineering, Saveetha School of Engineering, Saveetha ‎Institute of Medical and Technical Sciences, SIMATS 602 105 Chennai, India
    https://doi.org/10.14419/5fkr4104

    Received date: June 24, 2025

    Accepted date: September 4, 2025

    Published date: September 12, 2025

  • ASD; Non-ASD; Efficient Net; F1-Score.

  • Abstract

    This research evaluates the effectiveness of the Efficient Net model series (B0–B7) in detecting Autism Spectrum Disorder using facial image data. The findings indicate that the ‎deeper models attain better accuracy and more balanced classification results than the ‎shallower models. EfficientNetB3 and B7 achieve the top accuracy of 0.99, exhibiting excellent precision, ‎recall, and F1-scores for both ASD and non-ASD categories, emphasizing their effectiveness in reducing false ‎positives and false negatives. EfficientNetB2 and B5 also reach competitive accuracies of 0.98 and ‎‎0.97, offering dependable options with marginally lower complexity. Conversely, EfficientNetB4 achieves the ‎lowest accuracy of 0.88 because of poor recall in the non-ASD category, indicating restricted generalization. ‎The results affirm that more profound Efficient Net models, especially B3, B5, B6, and B7, excel in feature ‎extraction and classification for ASD prediction, providing a trustworthy structure for the early ‎and precise detection of the disorder.

  • References

    1. Vemuri, "Prediction of Autism Spectrum Disorder Using Efficient Net," in Proc. IEEE ICACCS, Mar. 2023.
    2. M. S. Alam, M. M. Rashid, and R. Roy, "Empirical Study of Autism Spectrum Disorder Diagnosis Using Facial Images by Improved Transfer Learning Approach," Bioengineering, vol. 9, no. 11, Nov. 2022, https://doi.org/10.3390/bioengineering9110710.
    3. T. S. Mian, "Efficient Net-based Transfer Learning Technique for Facial Autism Detection," Scalable Computing: Practice and Experience, vol. 24, no. 3, Sep. 2023, https://doi.org/10.12694/scpe.v24i3.2233.
    4. K. K. M. Rahman and M. M. Subashini, "Identification of Autism in Children Using Static Facial Features and Deep Neural Networks," Brain Sci-ences, vol. 12, no. 1, Jan. 2022, https://doi.org/10.3390/brainsci12010094.
    5. P. Reddy and A. J., "Diagnosis of Autism in Children Using Deep Learning Techniques by Analyzing Facial Features," Eng. Proc., vol. 5, Jan. 2024, https://doi.org/10.3390/engproc2023059198.
    6. J. Anjum, N. A. Hia, and A. Waziha, "Deep Learning-Based Feature Extraction from Children's Facial Images for Autism Spectrum Disorder De-tection," in Proc. ACM, May 2024, https://doi.org/10.1145/3660853.3660888.
    7. M. S. V. S. K. Narala, S. Vemuri, and C. Kattula, "A Comparative Study for Autism Spectrum Disorder Using EfficientNet, ResNet50," in Proc. IEEE ICCINS, Dec. 2023, https://doi.org/10.1109/ICCINS58907.2023.10450124.
    8. M. Tan and Q. V. Le, "EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks," arXiv preprint arXiv:1905.11946, May 2019.
    9. A. Bhise, S. Gawade, A. Pawar, A. Singh, and S. Shiravale, "Facial-Based Autism Spectrum Disorder Diagnosis Using Enhanced Deep Learning Models," in Proc. IEEE INC, Mar. 2024, https://doi.org/10.1109/InC460750.2024.10649230.
    10. C. Wang and J. Miao, "Deep Learning-Based Recognition of Autism Using Facial Datasets," in Proc. SPIE, https://doi.org/10.1117/12.3042402.
    11. S. L. K. Vinti, B. Panduri, V. Khedkar, M. B. Ketha, A. Lakshmanarao, and T. Srivalli, "Advanced Hybrid Transfer Learning Approaches for Au-tism Spectrum Disorder Detection Using Facial Features," in Proc. IEEE ICDICI, Nov. 2024.
    12. M. Sun, H. Dou, and J. Yan, "Efficient Transfer Learning via Joint Adaptation of Network Architecture and Weight," in Deep Learning for Vision Systems, Springer, 2020, pp. 381–398, https://doi.org/10.1007/978-3-030-58601-0_28.
    13. Y. B. Sree, O. Kumar, A. Sathvik, B. Sai, P. Rao, D. Sai, and H. Akshit, "Transfer Learning Based Autism Detection in Children," in Proc. IEEE INOCON, Mar. 2024, https://doi.org/10.1109/INOCON60754.2024.10511706.
    14. A. Singh, M. Laroia, A. K. S. Rawat, and K. R. Seeja, "Facial Feature Analysis for Autism Detection Using Deep Learning," in AI Applications in Healthcare, Springer, Jan. 2023, pp. 553–564, https://doi.org/10.1007/978-981-99-3315-0_41.
    15. B. Benaziz, A. Zitouni, and A. Amrouche, "Children Autism Detection Using Residual CNN and EfficientNetB1," in Proc. IEEE ICEEAT, Nov. 2023, https://doi.org/10.1109/ICEEAT60471.2023.10425848.
    16. R. Pilli, T. Goel, and R. Murugan, “Unveiling Alzheimer’s Disease through Brain Age Estimation Using Multi-Kernel Regression Network and Magnetic Resonance Imaging,” Comput. Methods Programs Biomed., vol. 261, p. 108617, Apr. 2025, https://doi.org/10.1016/j.cmpb.2025.108617.
    17. K. Dinesh, S. Kamatchi, and K. Mangaiyarkarasi, “Accuracy improvement for personality prediction using logistic regression in comparison with random forest algorithm,” in AIP Conf. Proc., vol. 3161, no. 1, p. 020269, 2024, https://doi.org/10.1063/5.0229463.
    18. A. Priya and P. S. Bharathi, “SE-ResNeXt-50-CNN: A deep learning model for lung cancer classification,” Appl. Soft Comput., vol. 171, 2025. https://doi.org/10.1016/j.asoc.2025.112696.
    19. M. R and M. Arunraj, “Enhancing Healthcare Insurance Cost Prediction with a Metaheuristic Optimization Approach: Comparative Analysis of Machine Learning Models,” in Proc. 2025 3rd Int. Conf. Inventive Comput. Informatics (ICICI), Bangalore, India, 2025, pp. 1225–1232. https://doi.org/10.1109/ICICI65870.2025.11069726.

  • Downloads

  • How to Cite

    Krishnamoorthy, V. ., Veeramani , T. ., Indirani , M. ., Sathishkumar, B. R. . ., Sasi , G. ., & Selvakumarasamy, K. . . (2025). Autism Spectrum Disorder Prediction from Facial Images Using Fine-Tuned Efficient Net B0–B7 ‎Architectures. International Journal of Basic and Applied Sciences, 14(5), 379-389. https://doi.org/10.14419/5fkr4104