Image Tracking

Image recognition is a more common form of presentation in AR applications, through the enhancement of the image function, after recognizing the picture, you can load a three-dimensional virtual content above the picture. Picture recognition is mainly used to build recognition of flat images, such as poster photos, books or product packaging.

Recognition map

Recognition map is a picture used in picture recognition to identify and trigger AR content, the more obvious the features of the picture used to do the recognition map the better the recognition effect.

Model decoupling

Model decoupling means that after loading the corresponding AR content in the picture or object recognition, the picture will be taken away or hidden, so that the recognition map disappears from the current camera field of view, while the model can still stay in the same place, so that the content can be detached from the recognition map and maintain the display.

Multi-card recognition

Multi-card recognition refers to the simultaneous recognition of multiple pictures, and can be done to recognize different pictures loaded with different AR content, such as "battle card" type of effect.

Face Tracking:

Face enhancement is a popular form of AR application in recent years, and the main application method is to detect faces through the face tracking technology of AR

. After recognizing the face according to the feature structure, the business model is loaded to the specified position of the user's face. Facial enhancement has been widely used in popular beauty cameras and short video applications.

It is worth noting that "face detection" is not exactly the same as "face recognition", in the application of facial tracking scenarios, face detection is not concerned about who is in the current scene, but more concerned about whether the object in the current scene is a "face". In the face tracking application scenario, face detection does not focus on who the person in the current scene is, but more on whether the object in the current scene is a "face", and as long as the detection result matches the features and contours of the face, the corresponding AR content can be loaded.

Facial Expression Capture

Facial Expression Capture allows for real-time tracking of facial expressions. Based on real-time expression tracking, it allows the AR virtual content we add to follow the changes in our facial expressions, realizing actions such as blinking and sticking out the tongue. It also allows users to customize emoji emoticons with dynamic expressions of their own faces.

3D Object Recognition / Model Recognition

Object Detection, also known as 3D Object Recognition and Model Recognition, is mainly used to record the spatial characteristics of a 3D object in the real world, and then the user scans the current environment and triggers the corresponding AR content when this object is detected.

Object recognition technology has a very wide range of application scenarios, such as combining historical artifacts, toys, industrial equipment, auto parts and other physical objects for AR display. Based on the object detection technology, it can make AR with more triggering possibilities. Surface, 360° tracking, which cannot be supported by picture recognition, can be realized based on object detection technology.

Hand gesture recognition

Hand gesture recognition tracking can output real-time information related to the hand appearing in the camera preview, including the type of gesture, the coordinates of the gesture box, the orientation of the palm, the judgment of the left and right hands, and the number of key nodes of the fingers.

Hand gesture recognition can be used in AR marketing interactions, short social videos, games, and other scenarios, making it easier and faster for users to interact with AR content.