Study of Different Deep Learning Approach with Explainable AI for Screening Patients with COVID-19 Symptoms: Using CT Scan and Chest X-ray Image Dataset

Multiple imaging types improve screening capacity: The study tested deep learning models on both CT scans and chest X-rays simultaneously, rather than relying on a single imaging method. This approach could help healthcare systems conduct larger-scale COVID-19 screening when testing kit availability is limited.

Explainability is essential for clinical trust: The researchers used LIME (Local Interpretable Model-agnostic Explanations) to make their AI model's decisions transparent, identifying which visual features the model relied on to distinguish COVID-19 patients from others. This transparency is critical for building trust between medical professionals and AI systems in clinical settings.

Model performance varied across different architectures: Eight different deep learning models were tested (including VGG16, ResNet50, and DenseNet201), suggesting that no single approach works best for this task. Healthcare organizations implementing AI screening tools should evaluate multiple models rather than assuming one solution fits all scenarios.

Interpretability addresses the "black box" problem: By explaining why the AI made specific predictions about patient images, the study demonstrates how explainable AI can help clinicians understand and validate model decisions rather than blindly accepting automated results—a key requirement for responsible deployment in healthcare.