A Proportion-Based Anatomical Landmark Model for Standardized Acupoint Localization Using Digital Body Surface Mapping

Abstract

Standardized acupoint localization is a longstanding challenge in acupuncture education and clinical practice, as traditional positioning methods rely heavily on textual descriptions, proportional measurements, and practitioner experience, which often results in variability and limited reproducibility. With the growing demand for objective, teachable, and clinically applicable standards, digital body surface mapping provides a promising pathway to support the standardization of acupuncture point localization while preserving the theoretical foundations of traditional Chinese medicine. In this study, we propose a Proportion-Based Anatomical Landmark Model (PB-ALM) for standardized acupoint localization using digital body surface images. The proposed model digitizes classical acupuncture principles—such as proportional measurement (cun) and anatomical landmark referencing—by defining acupoints as relative positions within a proportion-based coordinate system rather than fixed absolute coordinates. Stable body surface anatomical landmarks are used as references, while lightweight computer-assisted image analysis is employed only to support landmark identification and geometric mapping, without reliance on deep learning or complex artificial intelligence techniques. A digital acupoint localization assistance prototype was developed to demonstrate the practical applicability of the proposed framework in acupuncture teaching and clinical guidance. Experimental validation was performed on multiple commonly used limb acupoints by comparing PB-ALM–based localization results with expert manual annotations. The proposed method achieved a mean absolute localization error of 4.2 mm and an RMSE of 5.1 mm. Compared with manual proportional measurement, PB-ALM reduced inter-operator standard deviation from 4.5 mm to 3.0 mm and improved the intraclass correlation coefficient from 0.72 to 0.91, demonstrating enhanced consistency and reproducibility.