Image Analysis System
Kitami Institute of Technology × i-enter Corporation,Ltd.
Image Analysis System
Kitami Institute of Technology × i-enter Corporation,Ltd.
Image Analysis System
Non-contact AI image analysis system that does not restrict athletes' movements
We have developed a technology to estimate the posture of athletes by AI image analysis without attaching sensors to their bodies in collaboration with Kitami Institute of Technology.
Curling delivery (the action of throwing stones) requires the ability to control the body in millimeters and 0.1 seconds. The system aims to improve the physical skills of curlers and visualizes the behavior of the delivery form. It is also possible to compare with top curlers by combining models.
Development of skeletal detection learning model using virtual space
AI image analysis is performed by taking pictures of the skeleton of curling players at the time of delivery with one visible light camera on each side without attaching sensors to the player's body.
The acquired data can be analyzed with dedicated software, the skeletal position of each part can be detected, and a 3D model can be generated. This system utilizes Ienter's patented "posture estimation learning system", which can estimate the posture (skeleton) of a person in a real video from the learning data generated by a 3D model of the virtual space.
This technology has made it possible to generate a large amount of training data in the development of learning models for skeletal detection.
Contributing to the improvement of the physical skills of curlers
The image analysis system uses two visible light cameras to photograph objects and perform AI image analysis.
Since there is no need to attach sensors to the players, it is possible to measure natural behavior data without restricting the player's behavior.
In addition, it is now possible to check the skeleton from any angle in 360 degrees with 3D images. In the past, posture confirmation was mostly done visually and relied on rules of thumb, but by using the system developed this time, it is possible to analyze one's own posture visualized and consider it based on comparative data with other athletes.
Skeletal detection of objects for which real-world video data is difficult to obtain
The image analysis system is a technology that is also used by Japan's top curlers who have participated in the Olympic Games. In addition, at the "Argo Graphics Kitami Curling Hall", anyone can experience the technology.
Currently, the technology is introduced for the purpose of estimating the posture of curling players, but by using virtual space, it will be possible to support 2D and 3D, and it will be possible not only to detect human skeletons, but also to create posture learning data and skeleton detection of various animals.
In addition, it is expected to expand the range of support for skeletal detection for subjects for which it is difficult to obtain real video data.
Contact Us
Please feel free to contact us for consultation or questions.
IoT/AI/XR(VR・AR・MR)/RPA/Application Development
/Web System Development/Business System Development
/Cloud Integration/Design
/UX/Digital Marketing