ALTO

An open XML format maintained by the US Library of Congress that stores the output of OCR (optical character recognition) together with the layout information of the scanned pages: character positions, confidence scores, candidate alternative characters, block and line geometry, and reading order. ALTO is commonly paired with METS for digital-library deliverables and is used throughout the cultural-heritage sector. In accessibility contexts ALTO matters because preserving character-level coordinates lets downstream tools regenerate accessible derivatives (EPUB, DAISY) and refine OCR errors over time without losing earlier correction work. Extensions have been proposed to handle non-Latin scripts, right-to-left and vertical writing modes, and language-specific features such as ruby annotations.