An Explainable Hybrid AI Framework for Enhanced Tuberculosis and Symptom Detection

2025-10-21

AI Summary

In resource-constrained settings, tuberculosis (TB) early screening faces dual challenges: severe shortages of radiologists and scarcity of high-quality annotated data. To address these, we propose an interpretable hybrid AI framework integrating a teacher-student architecture, collaborative supervised learning (for three-class TB diagnosis and multi-label symptom identification), and self-supervised learning, enhanced by a joint optimization mechanism across dual supervised heads and a self-supervised head. Attention visualization confirms the model’s focus on clinically relevant anatomical regions, substantially improving clinical interpretability. Evaluated on chest X-ray data, our method achieves 98.85% accuracy in three-class TB classification and a macro-F1 score of 90.09% for multi-symptom detection—both significantly surpassing baseline methods. The framework balances high diagnostic accuracy, low dependency on labeled data, and clinical trustworthiness, offering a practical technical pathway for intelligent TB triage in primary healthcare settings.

Technology Category

• Computer Vision: Interpretability, Explainability, and Transparency
• Machine Learning: Transparent, Interpretable, Explainable ML
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Web Mining and Content Analysis: Large pretrained models with web data
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Tuberculosis remains a critical global health issue, particularly in resource-limited and remote areas. Early detection is vital for treatment, yet the lack of skilled radiologists underscores the need for artificial intelligence (AI)-driven screening tools. Developing reliable AI models is challenging due to the necessity for large, high-quality datasets, which are costly to obtain. To tackle this, we propose a teacher--student framework which enhances both disease and symptom detection on chest X-rays by integrating two supervised heads and a self-supervised head. Our model achieves an accuracy of 98.85% for distinguishing between COVID-19, tuberculosis, and normal cases, and a macro-F1 score of 90.09% for multilabel symptom detection, significantly outperforming baselines. The explainability assessments also show the model bases its predictions on relevant anatomical features, demonstrating promise for deployment in clinical screening and triage settings.

Problem

• Detecting tuberculosis and symptoms from chest X-rays
• Addressing lack of skilled radiologists in remote areas
• Overcoming limited datasets with hybrid AI framework

Innovation

• Teacher-student framework enhances disease and symptom detection
• Integrates supervised and self-supervised heads on X-rays
• Achieves high accuracy for tuberculosis and COVID-19