2015 · 21 citations · 5 references
ComputingHealth InformaticsAssistive TechnologyComputer VisionPhysical ActivityAction RecognitionMotion CapturePhysical Activity AssessmentPhysical FitnessPhysical Fitness AssessmentVirtual RealityActivity MonitoringCorrective Action AssessmentHuman-machine CollaborationPhysical Activity MonitoringPhysical Activity TrackingReal-time Processing
Our daily life is more convenient than ever before because of the great progress of science and technology. However, the body building or fitness are always ignored in modern daily life. In this paper, we propose a Virtual Personal Trainer to provide real time visually action guide and action assessment during the fitness time of users via the Microsoft Kinect Sensor. The user actions are captured by Kinect and compared to standard actions so as to give a fitness score representing how well users perform actions in real time. Users can see the skeletons of his/her own action and standard action in the screen together with videos. An action correctness guide will be displayed to guide the user to correct his/her actions. Users correct his/her actions to match the standard action in our pre-collected fitness action library. Thus our system can provide interpretable feedback on how people can improve their action. The whole system gives a new definition of the family gym in the future with the power of human-machine interactive.
5
Real-time human pose recognition in parts from single depth images
Jamie Shotton, Andrew Fitzgibbon, Mat Cook et al. · 2011 · 3.5K citations
Learning deformable action templates from cluttered videos
Benjamin Yao, Song‐Chun Zhu · 2009 · 66 citations
Fast 3D modeling in complex environments using a single Kinect sensor
Haosong Yue, Weihai Chen, Xingming Wu et al. · Optics and Lasers in Engineering · 2013 · 36 citations