FRX (Face tracking)
The technology to detect and track the presence of a human face in a digital video frame to enable FaceAR camera experiences.
Neural network that separates a user on the foreground from the background in a sequence of video frames to remove the background or replace it with another graphical image.
Face occlusion detection neural network that enables FRX performance with facial occlusions such as sunglasses, scarf, mask etc.
Neural network employed in different aspects of face filters, beautification and virtual makeovers to segment the skin and allow for its modification such as change the color and tone.
Neural network to detect and segment the image into hair, face and background to allow for real-time hair modification such as change the hair color.
Neural network to detect and segment the eye into iris, pupil and eyeball to allow for real-time eye modification such as recoloring, size adjustment, eyeshadows and eyelashes application.
Neural network to detect and segment the lips on the users face for real-time lips recoloring, lips size change in virtual makeup applications and beauty face filters.
Full body segmentation
Human body subtraction in full length in images and videos. The core of the background separation technology is a convolutional neural network which returns a binary output, tagging the image parts as either human or the background.
Real-time distance estimation between the face of the user and the device camera. Normally used to give the user feedback when he or she holds the device too far or too close to the face.
Algorithm detects micro-movements of the eye and its areas with subpixel accuracy in real-time. Based on that data, a vector of movement can be created. The technology makes it possible not only to “track” a person’s gaze but also to control a smartphone’s function with a gaze.
2+ faces detection
Algorithm allowing for masks application to several people simultaneously for more engaging group AR experiences. For the quality user experience, we generally don’t recommend supporting more than 3 faces on mobile devices due to limited computing capabilities.
Pulse (Heart rate)
Algorithm analyses fine patterns of the facial areas and their color variations within time to detect pulse frequency in real-time.
Mask on a picture from Camera Roll
Mask application on pre-recorded images the user uploads from the Camera Roll.
Mask on video from Camera Roll
Mask application on pre-recorded videos the user uploads from the Camera Roll.
Graphical camera effects and animations applied on pre-recorded videos.
Continuous photo editing
AR effect is processed in real-time on the image. E.g. beautification slider to control the face modification or "Before/After" slider.
Small AR scenarios enabled thought the user touches on the screen. AR objects or camera effects can change color and behaviour. Applied in FaceAR games or interactive face filters to increase engagement.
Small AR scenarios enabled through user facial expressions. The user can interact with effects or call them opening mouth, smiling, raising eyebrows or frowning. Applied in FaceAR games or interactive face filters to increase engagement.
Sound effects support in FaceAR experiences, e.g. add music to filters.
Acne removal is an algorithm to automatically remove acne by single taps on the selected area. Works in photos.
The 3rd party software allowing to change the tone of the user's voice, e.g. make the user talk like a robot, kid, male/female, etc.
Changing the size and proportions of the face, e.g. slim down the cheeks, nose or modify them for fun masks.
Skinned mesh animation
AR models look not static but moving, animated and transforming.
AR models behave like the real objects in the flow of real-world light and physics, e.g. support gravity or mirror the light with the camera rotates and user tilts.
Real-time or offline color correction of pre-recorded images, e.g. Instagram-like filters.
The digital representation of the surface of an AR object providing the sophisticated and life-like object representation.
Infuse a static image with dynamic qualities and explicit action to achieve an enhanced look and feel of the FaceAR video experience.
Face retouch and modification features including skin smooth, morphing, teeth whitening, eye’s modification and ect.
The fundamental actions of individual muscles or groups of muscles of the face that enable the AR mask to support user facial expressions, e.g. in emojis, avatars or full-face AR masks.