3D Human Motion Analysis to Detect Abnormal Events on Stairs

Abstract

Falls on the stairs are a common cause of accidental injury among the older adults. Understanding the mechanisms leading to such accidents may improve not only the prevention of falls, but also support independent living among elderly. Thus, a method to automatically detect falls and other abnormal events on stairs is presented and empirically validated. Automatic fall detection will also assist in data collection for environmental design improvements and fall prevention. Real-time 3D joint tracking information, provided by a Microsoft Kinect, is used to estimate the walking speed and to extract a set of features that encode human motion during stairway descent. Supervised learning algorithms, trained on manually labelled training data simulated in a home laboratory, obtained a high detection accuracy rate of ∼ 92% in leave-onesubject-out cross validation. In contrast with previous research, which identified visual tracking of the feet as the best indicator of dangerous activity, 3D motion of the hips is experimentally shown to be the most informative component in detecting abnormal events in the 3D tracking data provided by the Kinect.

Quantitative Mobility Assessment for Fall Risk Prediction in Dementia: A Systematic Review

“…Such a system in the home environment could monitor naturalistic movements remotely, and detect changes in response to interventions or therapies. Examples of emerging technologies include wearables such as accelerometers [29,51], vision-based data capture [52,53], or real-time location sensor networks [29]. Vision-based technologies blend into older adult living environments and track their movement without the need for wearables.…”

Evaluation of cupboard door sensors for improving activity recognition in the kitchen

“…Existing skeleton-based approaches have either used the full set of joints for general action recognition [15][16][17][18] or a subset chosen depending on the specific action/application [8,19] . In [19] , only hips, knees, ankles and feet joints were used for detecting abnormal events during stair descent. The method in [8] used feet joints along with the projection of hand and torso joints for evaluating musculoskeletal disorders on patients who suffer from Parkinson's Disease (PD).…”

A Survey of Applications and Human Motion Recognition with Microsoft Kinect

“…11,80,87,91,103,112,147 Bian et al presented an approach to detecting falls by extracting skeleton data from Kinect depth images based on the fast randomized decision forest algorithm. 11 This algorithm produces more accurate detection by properly rotating frames to match human orientation.…”

“…91 This method automatically estimates walking speed and extracts a set of features that encode human motion during stairway descent.…”