FirsthandHealth
medRxiv PreprintsInternational19 November 2026

Lightweight and Explainable AI for Chest Radiograph Classification in Low-Resource Settings: Cross-Dataset Generalization, Grad-CAM Verification, and Uncertainty-Aware Referral

This is an official announcement record

Firsthand records what medRxiv Preprints announced and links to the original. The wording below is theirs, not ours.

Chest radiography is the most common imaging test worldwide, yet models trained on one hospital's data often fail elsewhere, a critical problem for low-resource hospitals that cannot collect large local datasets. This paper presents a lightweight, explainable framework for chest radiograph classification, trained and evaluated entirely on CPU with public data. EfficientNet-B0 and ResNet-18 are trained on a 10,000-image subset of NIH ChestX-ray14 and evaluated without retraining on a paediatric pneumonia set, alongside a TorchXRayVision DenseNet121 pretrained baseline and a model trained direct
Source: medRxiv Preprints
Read the official announcement

Opens www.medrxiv.org

More from medRxiv Preprints

This content is for informational purposes only and is not medical advice. It is not intended to diagnose, treat, cure, or prevent any disease. Consult a healthcare professional before starting any supplement, treatment, or program: especially if you are pregnant, nursing, taking medication, or managing a health condition.