A Japanese tech company called NEXT-SYSTEM recently released VisionPose Single3D, a mark-free human skeleton detection software engine, for Unity-based platforms. Its technology is a “pose estimation engine” that uses AI and cameras to detect human skeletons from camera videos and still images. The engine doesn’t use any markers or depth sensors while detecting in real-time.
The engine of VisionPose Single3D has an array of applications in fields such as medical, sports and fitness, safety surveillance, entertainment, and gaming. The job of the AI is to detect 30 key points in 2D and 17 key points in 3D. These key points refer to areas such as the shoulders, knees, legs, arms, pelvis, nose, and others on the skeleton.
Since its entry into the tech world in 2002, NEXT-SYSTEM has been analyzing human behavior, especially in terms of motion through AI. The company also has been invested in developing ergonomics systems. With VisionPose Single3D, they bring MICHICO, a full-body real-time motion-capture application, that displays the movements of humans through a 3DCG character. This global venture is available for a 30-day free trial.
Let’s meet MICHICO
MICHICO replicates the human movement in real-time and the results will be displayed in a video format at around 30 FPS. The command “BodyAndColor with MICHICO” will allow the character to capture to track a person’s skeletal structure through a video or a photo.
With the inclusion of an application tool called VP Analyzer, the body movements are so precisely measured that MICHICO can easily perform varied movements automatically. NEXT-SYSTEM said that they hope “it will be utilized by various clients around the world, and that it will be of great use to them”.
Precise skeletal analysis
The software development kit that is part of the VisionPose environment enables it to detect human poses using a single camera. NEXT-SYSTEM says that this application will be extremely useful in the development of games and VR/AR tech. The key points on an adult skeleton structure are also precisely sized down to that of a baby. This, according to the company, will help the nursing sectors or the daycare sectors to keep track of babies.
The pose detection technology has also found a market in the fitness sector where age does not resist the flow of this technology. The engine estimates the current posture of the people under the scan. This estimation can be used for posture correction, accurate dance movements, and rehabilitation of the limbs based on the estimates. The ultimate goal is to help people adopt a healthy lifestyle by adopting the correct posture.
YouTube: Pose Estimation AI Engine “VisionPose”
Photo credit: The images used are owned by NEXT-SYSTEM and have been provided for press usage.