https://tech.ldas.jp/en/tags/japanese-ocr/Recent content in Japanese OCR on Digital Archive Systems Tech BlogHugoenTue, 31 Mar 2026 00:00:00 +0900https://tech.ldas.jp/en/posts/ndlocr-lite-python-integration/Tue, 31 Mar 2026 00:00:00 +0900https://tech.ldas.jp/en/posts/ndlocr-lite-python-integration/<p><a href="https://github.com/ndl-lab/ndlocr-lite">NDLOCR-Lite</a> is a Japanese OCR engine published by the National Diet Library of Japan. It extracts text from digitized images of books and periodicals using a combination of layout recognition (DEIM) and character recognition (PARSeq).</p> <p>It can be used via the <code>ndlocr-lite</code> CLI command, but when integrating it into a batch processing pipeline or handling recognition results programmatically, calling it from Python as a library is more practical.</p> <p>However, NDLOCR-Lite does not expose a public library API. The approach described here calls the internal entry point used by the CLI directly. This article documents how to do that.</p>