Why OCR of Acupuncture Classics Is Needed
Making the accumulated history of points, point combinations, needling, and moxibustion researchable
Acupuncture and moxibustion intervene in the body through the selection of acupuncture points and needling or moxibustion techniques. From antiquity to the present, practice has developed through a concrete historical accumulation of which points, combinations, and techniques were used for which presentations. For students and practitioners, tracing this accumulation back through the literature is essential to understanding and transmitting the techniques.
Acupuncture practice does not follow a single standardized methodology. Meridian therapy, Chinese acupuncture, modern acupuncture, traditional family lineages, and the particular techniques of different schools coexist. These schools do not necessarily rely on the same classical sources. Only by reading across works can one understand the historical background of one's points, combinations, and needling methods in depth.
Many acupuncture classics, however, are not available as searchable modern digital texts. Where works exist only as images, a point or disease name cannot be searched across the whole text, and preparing Japanese reading texts or modern Japanese translations requires substantial effort. This project converts these classics into machine-readable text to build a foundation for research, comparison, and organization. This is infrastructure for consulting classical literature as research, not for issuing clinical recommendations.
At the heart of acupuncture practice is the selection ofacupuncture points. For a given presentation, which points, combinations, and needling or moxibustion methods to use is a clinical choice with consequences for the outcome. Classical texts record relationships accumulated over time among presentations, points, combinations, and techniques. Questions include:
A single textbook cannot answer these questions. Multiple acupuncture classics must be consulted close to the original, then compared and organized through Japanese reading and modern translation. This is research reference to classical materials, not a clinical recommendation . Each acupuncturist makes clinical judgments based on their school and experience; the project prepares the reference foundation.
OCR (Optical Character Recognition) converts classical pages stored as images into text that can be searched, compared, and copied. OCR of old books presents difficulties distinct from modern Japanese OCR:
The project therefore uses an OCR engine developed for classical books. OCR is a machine-processing stage; comparison with originals and critical editing must proceed separately in stages.
The project selected Sun Simiao's Tang-period Beiji Qianjin Yaofang (Qianjin Fang) as a starting draft for preparing acupuncture-classics OCR. An encyclopedic work in East Asian medical history covering medicinals, formulas, acupuncture, and disease presentations, it is a major source. Using Qianjin Fang RB00003825 in Kyoto University's Rare Materials Digital Archive (Fujikawa Collection), image acquisition and OCR were completed for every page.
The OCR engine used was NDLkotenOCR-Lite , a classical-book OCR engine released by the National Diet Library, with support for vertical texts in Japanese woodblock editions and Chinese books, including pages with variant kana and characters.
“Completion” of all-page OCR means completion of machine processing, not completion of translation or critical editing. The project distinguishes the following stages:
| Stage | Meaning | Current Qianjin Fang Status |
|---|---|---|
| Uncorrected OCR Output | Machine-read text after OCR processing; uncorrected and potentially containing misreadings, omissions, and misplaced line breaks. | Complete |
| TEXT_CLEANED | Obvious OCR errors corrected, with paragraph and chapter structures organized. | Not Started |
| Text Verified | A critically edited state following comparison with original images and expert verification. | Not Started |
At the stage described here, the project holds uncorrected OCR output, not a verified text. It is inappropriate to use this Qianjin Fang text directly as the basis for clinical judgment or scholarly quotation.
Qianjin Fang is the starting draft for acupuncture-classics OCR. The plan is to add other important works progressively, including those below, after Qianjin Fang OCR quality evaluation reaches an appropriate stage.
Electronic texts of acupuncture classics circulate from several sources. This project explicitly distinguishes sources rather than conflating them.
Because their sources and preparation processes differ, these texts must be distinguished when comparing or quoting them.
Text-specific gateways are organized in Classical Texts of East Asian Medicine . Free introductions to Qianjin Fang, Waitai Miyao, Zhubing Yuanhou Lun, and other works will be expanded progressively.
Acupuncture-classics OCR is important but still under preparation. The long-term aim is an environment for cross-searching and consulting relationships among points, combinations, presentations, needling, and moxibustion across multiple classics, allowing practitioners and researchers to examine the classical background of their techniques. This is research infrastructure, not automated treatment advice. It will provide foundational data for understanding MLMN's L5(channel layer) through comparison across texts.