Noninvasive methods enabling measurement of shoulder bone positions are paramount in clinical and ergonomics applications. In this study, the acromion marker cluster (AMC) method is assessed in comparison with a model-based approach allowing scapula tracking from low-dose biplanar radiograph images. Six healthy male subjects participated in this study. Data acquisition was performed for 6 arm abduction positions (0°, 45°, 90°, 120°, 150°, 180°). Scapula rotations were calculated using the coordinate systems and angle sequence was defined by the ISB. The comparison analysis was based on root mean square error (RMSE) calculation and nonparametric statistical tests. RMSE remained under 8° for 0° to 90° arm abduction and under 13.5° for 0° to 180° abduction; no significant differences were found between the 2 methods. Compared with previous works, an improved accuracy of the AMC approach at high arm abduction positions was obtained. This could be explained by the different sources of data used as the "gold standard."