Abstract :

This Project Presents A Novel Generative Artificial Intelligence (AI)-based Framework For The Early Detection Of Autism Spectrum Disorder (ASD) Using Facial Image Analysis. Traditional Diagnostic Methods For ASD Primarily Rely On Behavioral Assessments And Clinical Observations, Which Are Often Subjective, Time-consuming, And Dependent On Expert Availability. These Limitations Can Lead To Delayed Diagnosis And Reduced Effectiveness Of Early Intervention Strategies. To Address These Challenges, The Proposed System Leverages Advanced Deep Learning Techniques, Specifically Convolutional Neural Networks (CNNs) And Generative Adversarial Networks (GANs). The CNN Model Is Utilized To Automatically Extract Meaningful Facial Features And Subtle Visual Patterns Associated With ASD, While The GAN Model Generates Synthetic Facial Images To Enhance Dataset Size And Diversity. This Combination Improves Model Generalization, Reduces Overfitting, And Enhances Classification Accuracy. The System Follows A Structured Pipeline Including Data Collection, Preprocessing, Feature Extraction, Data Augmentation, Classification, And Performance Evaluation. The Model Classifies Facial Images Into ASD And Non-ASD Categories And Is Evaluated Using Metrics Such As Accuracy, Precision, Recall, And F1-score. Experimental Results Demonstrate Promising Performance, Achieving Accuracy Levels Of Approximately 85% To 92%, Indicating The Effectiveness Of The Proposed Approach. This Framework Provides A Non-invasive, Efficient, And Scalable Decision-support Tool For Early ASD Screening. Although Not Intended To Replace Clinical Diagnosis, It Assists Healthcare Professionals In Making Faster And More Objective Assessments. The Study Also Highlights The Potential Of Integrating Generative AI In Medical Imaging Applications, Paving The Way For Future Research Involving Multimodal Data Such As Speech And Behavioral Analysis.

Published:

24-4-2026

Issue:

Vol. 26 No. 4 (2026)

Page Nos:

2900 - 2908

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite