An Empirical Study of Monocular Human Body Measurement Under Weak Calibration

TL;DR

This paper tackles the challenge of estimating major human body measurements from monocular RGB images under weak calibration. It compares three strategies—landmark-based anthropometric baseline, pose-driven regression, and object-calibrated silhouette estimation—within a unified pipeline and semi-constrained capture setup on consumer hardware. The findings reveal a clear calibration-robustness trade-off: stronger calibration improves depth-sensitive and circumferential estimates, particularly waist, while purely pose-based methods remain susceptible to scale drift and viewpoint variation. The work provides a practical empirical reference for designing lightweight monocular body-measurement systems suitable for deployment on consumer devices, highlighting clothing, lighting, and silhouette fidelity as key constraints.

Abstract

Estimating human body measurements from monocular RGB imagery remains challenging due to scale ambiguity, viewpoint sensitivity, and the absence of explicit depth information. This work presents a systematic empirical study of three weakly calibrated monocular strategies: landmark-based geometry, pose-driven regression, and object-calibrated silhouettes, evaluated under semi-constrained conditions using consumer-grade cameras. Rather than pursuing state-of-the-art accuracy, the study analyzes how differing calibration assumptions influence measurement behavior, robustness, and failure modes across varied body types. The results reveal a clear trade-off between user effort during calibration and the stability of resulting circumferential quantities. This paper serves as an empirical design reference for lightweight monocular human measurement systems intended for deployment on consumer devices.

Figures (2)

• Figure 1: Overview of the evaluation pipeline
• Figure 2: Frontal and lateral person silhouettes extracted under fitted clothing conditions, where clothing preserves body contour.