The Avatars from Images feature allows you to easily generate an accurate 3D model of yourself using two different modes: Classic mode ****and the experimental Instant Avatar mode. Each mode has its own strengths and applications, depending on your needs.
Classic mode is designed to provide accurate results for form-fitting clothes. It uses a technique called silhouette fitting, which generates a 3D model based on the contour of your body in the image. This mode is ideal for avatars with an accurate representation of your form, particularly when wearing tight-fitting clothing.
Example of the silhouette fitting process using classic mode.
Instant Avatar Mode is an exciting experimental feature designed to adapt to a variety of clothing styles, including loose or non-form fitting garments.
Powered by BEDLAM, an innovative neural network-based method for creating 3D avatars, this mode offers a less restrictive approach, allowing you to wear your preferred outfit and not be confined to a specific environment. However, please keep in mind that it might produce slightly less accurate results overall than Classic mode approach.
Source image
Instant Avatar result
Following you will find which image(s) work best for the classic and instant avatar mode, as well as general best practices in order to achieve high accuracy.
20230726_Best Practices_Avatars from Images.pdf
In Classic mode, you can upload one photo of a person wearing form-fitted clothes. For optimal results, choose high-contrast settings (between the subject and environment), an easy upright pose without intersections of body and hands or arms, and have the subject wear form-fitting clothing. Make sure the subject's full body is visible in each photo, with longer hair tied up, and position the camera to face the subject directly without tilting.
Example of the ideal pose for classic mode.
Instant Avatar mode requires a single photo as well. While it is designed to effectively capture various clothing styles, including loose or non-form fitting garments, and accommodate non-standard poses, it is still recommended to maintain a neutral pose with hands and feet visible for optimal results.