You can enter one or more measurements, and our algorithm will use those measurements to predict the rest for your body. This way, you can create a customized avatar that looks just like you! Whether you enter all of your measurements or just a few, our algorithm will use the information you provide to create a realistic and accurate representation of your body. But how does that work?
Our measurements follow the **ISO 8559-1:2017(E)** standard.
| UI label | ISO Standard name | API measurement key |
|---|---|---|
| Height | Height | Height |
| Weight | not included | Weight |
| Chest | Bust girth | Bust_girth |
| Waist | Waist girth | Waist_girth |
| Hips | Top hip girth | Top_hip_girth |
| Inseam | Inside leg height | Inside_leg_height |
| Shoulder length | Shoulder length | Shoulder_length |
| Upper arm length | Upper arm length | Upper_arm_length |
| Lower arm length | Lower arm length | Lower_arm_length |
| Neck base girth | Neck base girth | Neck_base_girth |
| Thigh girth | Thigh girth | Thigh_girth |
| Ankle girth | Ankle girth | Ankle_girth |
| Armscye girth | Armscye girth | Armscye_girth |
For more information about the measurements and their corresponding ISO landmarks and SMPL vertex, please check out our measurements manual.
20231107_Measurements-Manual.pdf
Meshcapade Me is powered by the SMPL model, a 3D human body model that was developed using a collection of 1.2 million body scans. This wealth of body scans was gathered from three primary datasets: scans performed at the Max Planck Institute for Perceiving Systems, the CAESAR dataset on the human body shape, and the Size USA dataset. With access to these rich and extensive datasets, we could learn more about the human body and its proportions than if we spent the rest of our lives meeting and measuring individuals. Using these datasets, we train a machine learning model which can take a few measurements as input and predict the rest of the measurements.
While many bodies will be well-represented by our predictive measurements, some won’t be - and that’s perfectly normal! All bodies are different in subtle and beautiful ways. Our predicted measurements draw on millions of user data points and are often accurate, but not always - if your weight or proportions vary from the predicted model, don't worry. This is how machines learn and is in no way a statement of value or medical advice. You're perfect just as you are - and with your help, the avatar tech can become even better.
Because while we are drawing on an immense amount of data, the majority of our non-user scans were collected in the United States and Europe, and don't necessarily capture the full cross-section of humanity that we strive to serve in future. That’s why we need the active participation of Meshcapade users - by opting in to have your anonymized measurement data retained within our growing data sets, we can continue to improve our machine learning and model an ever-broader spectrum of bodies.
You have the option to download an avatar’s measurements as a JSON. This feature is available for all avatars, whether they were created from measurements or images.
<aside> <img src="/icons/info-alternate_gray.svg" alt="/icons/info-alternate_gray.svg" width="40px" /> Note: Avatars created before the v1.6 release (13 April 2023) do not currently support this feature. However, this functionality will be added in a future release.
</aside>
From the Editor, click on the download button
Select Export measurements from the drop down menu.