Postural Assessment System based on Kinect
Computer vision, Poor Postural Detection, Point Cloud
Computer technology related activities usually requires long hours in a row in the same position (sitting), which is generally established improperly, causing poor posture. There are recommendations and exercises to avoid that and even physical therapy to try to remedy the damage. But these practices are usually used when there is already a postural problem and the person has pain. This research carries out a study of postural problems that may be arising with bad posture and develops a system that identifies in real-time if the individual is in an inappropriate position, based on a calibration of the initial state (proper), using Microsoft kinect Sensor and open source point cloud library library (PCL) to handle the cloud of points acquired. Each frame a comparisons between the point clouds extracted from Kinect are performed in order to verify the posture. If it is not appropriate, the user is warned of the poor posture.