Compact Multimodal Language Models as Robust OCR Alternatives for Noisy Textual Clinical Reports

TL;DR

This paper evaluates compact multimodal language models as robust, privacy-preserving OCR alternatives for noisy, regionally inflected clinical reports, demonstrating superior accuracy over traditional OCR methods.

Contribution

It introduces and benchmarks compact multimodal models for noisy medical text transcription, highlighting their robustness and adaptability in healthcare digitization.

Findings

Multimodal models outperform classical OCR in noisy conditions.

They maintain high accuracy with degraded images.

Despite higher computational costs, they are suitable for on-premises healthcare use.

Abstract

Digitization of medical records often relies on smartphone photographs of printed reports, producing images degraded by blur, shadows, and other noise. Conventional OCR systems, optimized for clean scans, perform poorly under such real-world conditions. This study evaluates compact multimodal language models as privacy-preserving alternatives for transcribing noisy clinical documents. Using obstetric ultrasound reports written in regionally inflected medical English common to Indian healthcare settings, we compare eight systems in terms of transcription accuracy, noise sensitivity, numeric accuracy, and computational efficiency. Compact multimodal models consistently outperform both classical and neural OCR pipelines. Despite higher computational costs, their robustness and linguistic adaptability position them as viable candidates for on-premises healthcare digitization.