SMPL is a whole family of 3D body, face, hand and foot models that are learned from hundreds of thousands of detailed 3D scans.

SMPL uses machine learning to capture the statistics of 3D body shape across the population.

No 3D body model is more accurate or even comes close.

For portability, SMPL is built on industry-standard linear blend skinning (LBS). But LBS is not very realistic.

We use machine learning to correct the limitations of LBS and add soft tissue deformations to the body, bringing lifelike details to animations.

Detailed 3D body shape is controlled with only 10-300 numbers depending on the accuracy you need.

Because it knows about real people, you can create accurate 3D humans from very little data.

A few measurements or an image is enough.

Gender? Your body, your choice.

SMPL comes in female, male, and non-binary versions. The non-binary version can represent most people. Nobody has to choose or disclose a gender.

SMPL is compatible with every game engine and graphics package by design.

We started with compatibility and then used machine learning to create the most accurate avatar technology that is metaverse ready.

Why are so many companies and research labs using SMPL?

SMPL is the de facto standard. There are more resources, code, and datasets available for SMPL than any other 3D body model.