Inter-method agreement between manual and 3D anthropometric measurements: an ICC, Bland–Altman, and RMSE analysis

Abstract Background Accurate anthropometric measurements are fundamental for apparel engineering, ergonomics, healthcare, and digital human modeling. Conventional manual anthropometry remains the reference standard; however, three-dimensional (3D) body-scanning technology offers a rapid, non-contact alternative with the potential to improve measurement efficiency and reproducibility. Comparative validation studies are required to determine the reliability and agreement of 3D body scanning relative to traditional manual methods. Objective This study aimed to compare conventional manual anthropometric measurements with three-dimensional (3D) body-scanning measurements among young adult Bangladeshi males by evaluating measurement differences, reliability, agreement, and practical applicability. Methods A comparative cross-sectional study was conducted involving 300 healthy Bangladeshi males aged 20–25 years, recruited from all eight administrative divisions of Bangladesh using proportional representation. Anthropometric measurements, including stature, chest circumference, waist circumference, hip circumference, shoulder breadth, arm length, thigh circumference, and inseam length, were collected under standardized laboratory conditions using both manual anthropometric techniques and a structured-light 3D body scanner. Descriptive statistics, paired-sample t -tests, Intraclass Correlation Coefficients (ICC), Bland–Altman agreement analysis, Root Mean Square Error (RMSE), and Technical Error of Measurement (TEM) were used for statistical evaluation. Results Manual and 3D anthropometric measurements demonstrated statistically significant differences for several body dimensions ( p < 0.05); however, the observed effect sizes were small (Cohen's d = 0.15–0.29), indicating limited practical significance. The mean differences between methods ranged from 0.30 to 0.80 cm, with minor systematic bias observed primarily in circumferential measurements. Excellent agreement and reliability were demonstrated across all anthropometric variables, with ICC values ranging from 0.949 to 0.982. Bland–Altman analysis showed narrow 95% limits of agreement within acceptable anthropometric tolerance levels. Measurement errors remained low, with RMSE values ranging from 1.00 to 2.20 cm and low TEM values, confirming high measurement precision. Compared with manual anthropometry, the 3D scanning system substantially reduced measurement time while maintaining comparable measurement accuracy. Conclusion Three-dimensional body scanning provides a reliable, accurate, and operationally efficient alternative to conventional manual anthropometric measurement for young adult Bangladeshi males. The high level of agreement between the two methods supports the application of 3D body-scanning technology in apparel sizing, ergonomic product design, digital human modeling, and large-scale anthropometric surveys, while significantly improving measurement efficiency.