Portable Medical Imaging in Modern Healthcare: Fundamentals, AI-Based Taxonomy, Image Quality, and Open Challenges

A new academic review provides a systematic roadmap for using artificial intelligence to solve the critical image quality problems plaguing portable medical imaging (PMI). Published on arXiv by researchers Yassine Habchi, Hamza Kheddar, and colleagues, the paper focuses on modalities like portable CT, MRI, and ultrasound, which are vital for point-of-care diagnosis in rural and emergency settings but suffer from severe degradation due to motion artifacts, hardware limitations, and unstable acquisition conditions.

The core contribution is a novel, quality-centered taxonomy of AI methods designed specifically for PMI challenges. It categorizes approaches across machine learning, deep learning, transfer learning, and Transformer-based models, detailing their application in image enhancement, reconstruction, quality assessment, and disease detection. Unlike previous surveys, this work explicitly connects technical AI robustness with clinical usability, analyzing devices, sensing pipelines, modality-specific distortions, and available datasets to bridge the gap between lab research and real-world deployment.

Finally, the review identifies key research gaps and outlines future directions, pushing for more reliable, interpretable, and clinically deployable AI systems. This structured analysis is a significant resource for engineers and clinicians aiming to build trustworthy portable imaging tools that can deliver accurate diagnoses outside traditional hospital infrastructure.