AI – Behavior Identification

Timely Comparison and Identification of Behaviors and Motions

YUAN QDEEP SDK is empowered with the technology of human keypoint detection. By quickly locating and returning the 3D correspondence of each human keypoint, QDEEP SDK precisely detects human skeleton, behavioral modes and motions.
Based on human skeletal points, after detecting human postures and positions of different keypoints, QDEEP SDK will utilize the information of body movements to compare the resemblance of skeletal points and evaluate consistency of movements. The tech can be applied in behavior recognitions, classification of different motions, detection of abnormal movements and events, augmented reality (AR) etc.
Furthermore, behavior detection and analysis can be applied to smart classroom. With data collection, the tech records students’ various behaviors in class and understand their feedbacks and performance, lecturers can timely adjust the content and provide aids according to the taken data.
On the application of domestic care system, behavior detection and analysis serve as an efficient assistance to ease the problem of understaffing issue of current medical caring system, by providing timely detection of abnormal behaviors such as collapsing or crying for help.

Technical Advantages

Keypoints detection & Unrestrained by Number of People

18 Human keypoints can be acquired from the whole body, including eyes, ears, nose, neck, shoulder, hands, elbows, waist, knees and feet etc.

Simultaneous detections of keypoints of crowd of multitude, unrestrained by number of people and applicable in different scenes.

Path Tracking & Auto Labeling

QDEEP SDK acquires 3D XYZ correspondences and provide the information of relative relations between scenes and postures.

With YUAN Labeling software, users can define certain movements and behaviors to adjust training content according to their needs.

Examples for Application Scenarios

Smart Classroom

Augmented Reality

Long Term Domestic Care

Motion Analysis