Lost in OCR-Translation: Pixel-based Text Reflow to the Rescue: Magnification of Archival Raster Image Documents in the Browser without Horizontal Scrolling

Abstract

Low-vision users have been accustomed to viewing accessible electronic documents with reflow and responsive layouts that adapt to the user’s text size requirements. However, mainstream technologies do not work optimally with image-based documents scanned from paper as well as documents that do not adhere to accessibility guidelines. Users are often dependent on viewing magnifications of such documents by scrolling in two-dimensions which most users find confusing. This paper discusses the capabilities of key state-of-the-art technologies and describes a browser-based document magnification implementation that reflow document contents at the pixel-level to prevent horizontal scrolling. The implementation provides low-vision users a practical alternative for simplified access to the content of archival documents with a different view than the state-of-the-art technologies. The paper also discusses practical and operational issues as well as unresolved challenges.