3D human pose recognition for home monitoring of elderly

Abstract

A toolbox for the automatic monitoring of elderly in a nursing home or in the natural home environment is proposed. Rather than monitoring vital signs or other biomedical parameters, the toolbox is focussed on the monitoring of activity patterns and changes therein. Activity information is derived from visual information using image processing algorithms. The visual information is acquired using 3D camera technology. Besides a traditional visual image, 3D cameras also provide highly accurate depth information. The 3D position of the subject is derived and serves as the primary information source for the different components in the toolbox.

Fall Detection from Depth Map Video Sequences

“…Falls can be detected during the post-fall phase when the person is motionless on the ground just after the fall [7]. The centroid height of the person relative to the ground has been used to detect an abnormal position near the floor [5]. This method works well when the silhouette is not occluded by furniture in the room.…”

“…The centroid height of the person relative to the ground is an efficient method for fall detection [5]. By pursuing this idea, we will simplify this method with several improvements and will demonstrate that a low cost depth map system can efficiently detect falls using this information.…”

“…A TOF camera provides more accurate depth images than a stereovision system, but it is very expensive and currently limited to low image resolution (e.g. image size of 176x144 pixels in [5]). …”

“…One solution is to reconstruct the total volume information in the scene, for example with a visual hull [2,3], but this is expensive and difficult to set up (multiple calibrated and synchronized cameras are needed). Other works use partial volume information [5] obtained from a Time-of-Flight sensor which returns precise depth images.…”

“…This method works well when the silhouette is not occluded by furniture in the room. In [5], the camera localization relative to the ground was needed. Here this information is not necessary, since we just need to compute the distance from a point (body centroid) to a plane (floor).…”