Exercise training evaluation system and method based on cooperation of event camera and visual camera

Abstract

The invention discloses an exercise training evaluation system and method based on cooperation of an event camera and a visual camera, is applied to the field of computer vision, and aims to solve the problems that an existing training evaluation system based on an optical camera cannot capture a rapid motion track and limb shape, cannot effectively evaluate the motion quality, cannot effectively evaluate the training quality, and cannot provide effective training guidance. According to the method, continuous human motion events are provided in an event stream form, millisecond-level motion response can be achieved, the method is not affected by the motion blurring effect of a high-speed moving object, a higher dynamic range can be provided, and more effective motion training evaluation can be provided under the scenes of strong light, backlight and violent light and shade change. Through cooperation of the visual camera and the event camera, not only can the accuracy of the rapid action be evaluated, but also standard action teaching and training can be carried out in a visual scene interaction form.

Description

technical field [0001] The invention belongs to the field of computer vision, and in particular relates to a technology for fast motion recognition and quality evaluation by using an event camera and a depth camera. Background technique [0002] The visual camera based on the optical sensor collects the accumulation of pixel brightness values ​​over a period of time, which is limited by the light sensitivity and data reading rate, the frame rate is very limited, and it is easily affected by ambient light, high hardware performance requirements, high energy consumption, and delay. At the same time, there are problems of motion blur and long response time for high-speed moving objects, which are not ideal for action recognition and evaluation. The event camera is a special camera that is only sensitive to changes in pixel brightness. It is a new type of neuromorphic vision sensor. Its pulse neuromorphic system has asynchronous, low power consumption, low computing throughput, ...

AI Technical Summary

Problems solved by technology

Due to the fast and complex movements of competitive sports (requiring the coordinated participation of all parts of the body), the training evaluation system based on optical cameras cannot capture the fast movement trajectory and body shape, cannot effectively evaluate the quality of movements, and cannot provide effective training guidance

Images