We have multiple neural networks working together. Face tracking (FRX) detects the presence of a human face. Then the rest of them do their job: eye segmentation lips segmentation, skin segmentation, etc. As a result, each element of the face gets the appropriate AR filter, perfectly mimicking a cosmetic product.
This is what the users can try:
Foundation, Eyelashes, Eyeliner, Lipstick, Eyeshadow, Blusher, Highlighter, Contouring.

Realistic, high-definition hair coloring that works on both natural shades and wild, outlandish ones. A neural network segments the image into face, hair, and background and other algorithms allow distinguishing strands. A user can check out both single-color schemes and the ones that have strand highlights or multiple shades.

The system creates a 3D model of the user’s face. Then, making changes is as simple as moving the model’s vertices.
This is what can get morphed:
Face shape, Nose, Eyes, Lips, Chin, etc.

A multifaceted feature that helps users look their best at any time. Several neural networks, trained to segment and retouch certain parts of the face, ensure the quality of picture, similar to that of a celebrity photoshoot.
This feature includes the following filters:
Skin smoothing, Neck smoothing, Acne removal, Teeth whitening, Eye bag removal, “Expressive eyes”.

Preprogrammed or interactive AR scenarios (filters) that users can launch with facial expressions. The system detects landmarks (points on the face) and triggers the appropriate actions once once these landmarks take certain positions (e.g. mouth opens).
The following triggers are available:
Mouth open, Smile, Eye brows, Eyes (open/close).

Banuba Face AR SDK lets you add stickers, filters, celebrity likenesses, animal faces, and other effects while filming a clip or in post-production. This is done by placing effect files (snowflakes, flames, etc.) on top of the video.

Let the users hide the background for privacy or replace it with something wild and cool for entertainment. This is done by a neural network that separates the user in the foreground from the rest of the frame. Then the background can be changed. Such filters are popular in the video conferencing apps, Snapchat competitors, online streaming apps, etc.

Animated 3D avatars that copy the users’ facial expressions. To do so, the system tracks certain points on the user’s face and calculates their position on the avatar accordingly.
The following things can be tracked:
Eyebrow position (incl. tracking separate eyebrows), Mouth position, Cheek puffing, Jaw movement, Nose position.

Decrease returns and increase customer satisfaction by letting people try your products on their smartphones. The technology allows designing custom effects that perfectly replicate real-world items.
These objects can be virtually tried on:
Glasses, Jewelry (necklaces, earrings), Piercings, Hats.

Give players an option to put on their “game faces” with cool 3D masks, or use facial expressions for game controls. You can use JavaScript to make effects trigger by opening mouth, moving eyebrows, tapping the smartphone screen, etc. This allows creating fun casual games where users can literally eat virtual objects or control their character with their face.

Dynamic, transforming AR objects that make any video look magical, impressive, and very shareable. The system recognizes the user’s face and makes the effects stick to it.

Color correction for entire videos, similar to what Instagram and TikTok offer. This includes both simple filters (black-and-white, sepia, etc.) and more outlandish color schemes.