A research group consisting of assistant professor Yukitaka Yamamoto, graduate student Yuto Ishige, professor Yutaka Takemura of Tokyo University of Science, and professor Yasushi Shimatani of Prefectural University of Hiroshima, is working on pose estimation from one RGB camera image and one or two small inertial measurements. We have succeeded in developing a new method to analyze gait easily and accurately by combining the information obtained from the units.
While gait analysis is used in various clinical settings, including rehabilitation, it requires a large number of markers and large-scale measuring equipment, resulting in high costs.Since 3D motion capture, which is a representative measurement tool, and motion capture using inertial measurement devices have the same problems, a simpler alternative method has been sought.
On the other hand, in this research, we proposed a new gait analysis method that combines a human posture estimation algorithm that estimates the joint points of the human body from camera images and information obtained from inertial measurement units attached to the feet.Only one RGB camera, which is also used in digital cameras, and one or two small inertial measurement units are used for analysis, so no specialized knowledge or large-scale equipment is required, and anyone can use it. Gait analysis can be easily performed.
In a comparative demonstration experiment of this method, it was confirmed that it is possible to measure many parameters related to the entire walking, such as walking speed, stride length, and walking cycle, with a high accuracy equal to or higher than that of the conventional method.The problem with camera-based markerless systems using conventional posture estimation algorithms is that the measurement error of joint angles during walking, especially ankle joint angles, is large. It was also found that it can be measured more accurately than the method.
The results of this research will not only be useful in the field of medical care and welfare, but by making it possible to easily perform high-precision motion analysis, it is possible to expect applications in a wide range of fields such as sports, robots, and computer graphics. .
Paper information:[Scientific Reports] Verifcation of gait analysis method fusing camera-based pose estimation and an IMU sensor in various gait conditions