Human Mesh Modeling for Anny Body

TL;DR

The paper tackles the limitations of scan-dependent and demographically narrow 3D human body models by introducing Anny, a differentiable, scan-free model grounded in MakeHuman anthropometrics. Anny uses continuous phenotype parameters to control a set of blendshapes, calibrated to WHO growth statistics, enabling diverse morphologies from infancy to old age within a single model. It introduces Anny-One, a large synthetic dataset (800k images) to train and evaluate Human Mesh Recovery (HMR) methods, showing that Anny-based training can achieve state-of-the-art performance on standard benchmarks even without real scans. The work provides comprehensive interoperability with existing models, detailed shape statistics, and demonstrates strong HMR results in both single- and multi-person scenarios, making Anny a practical, privacy-friendly foundation for wide-ranging 3D human modeling tasks.

Abstract

Parametric body models provide the structural basis for many human-centric tasks, yet existing models often rely on costly 3D scans and learned shape spaces that are proprietary and demographically narrow. We introduce Anny, a simple, fully differentiable, and scan-free human body model grounded in anthropometric knowledge from the MakeHuman community. Anny defines a continuous, interpretable shape space, where phenotype parameters (e.g. gender, age, height, weight) control blendshapes spanning a wide range of human forms--across ages (from infants to elders), body types, and proportions. Calibrated using WHO population statistics, it provides realistic and demographically grounded human shape variation within a single unified model. Thanks to its openness and semantic control, Anny serves as a versatile foundation for 3D human modeling--supporting millimeter-accurate scan fitting, controlled synthetic data generation, and Human Mesh Recovery (HMR). We further introduce Anny-One, a collection of 800k photorealistic images generated with Anny, showing that despite its simplicity, HMR models trained with Anny can match the performance of those trained with scan-based body models. The Anny body model and its code are released under the Apache 2.0 license, making Anny an accessible foundation for human-centric 3D modeling.

Figures (9)

• Figure 1: Anny is a unified, open and interpretable human parametric body model aiming to capture the diversity of human shapes and ages, from infants to elders.
• Figure 2: Shape parametrization is implemented using piece-wise multi-linear interpolation between prototypical shapes (illustration with age).
• Figure 3: Example of local morphological variations covered by the model.
• Figure 4: Mesh and skeleton of Anny. Left: default mesh and skeleton, modeling two different morphologies. Middle: coarser mesh (1,229 vertices) using only a skeleton subset. Right: use of the same skeleton as a Mixamo mixamo character for pose re-targeting.
• Figure 5: Model calibration. Calibration of Anny morphological shape distribution with WHO Child Growth standards who2006growth (boys).