A Posture Recognition-Based Fall Detection System for Monitoring an Elderly Person in a Smart Home Environment

By Miao Yu, Adel Rhuma, Syed Mohsen Naqvi, Liang Wang, Jonathon A. Chambers. IEEE Transactions on Information Technology in Biomedicine, 2012

322 citations recorded by the source corpus.

We propose a novel computer vision-based fall detection system for monitoring an elderly person in a home care application. Background subtraction is applied to extract the foreground human body and the result is improved by using certain postprocessing. Information from ellipse fitting and a projection histogram along the axes of the ellipse is used as the features for distinguishing different postures of the human. These features are then fed into a directed acyclic graph support vector machine for posture classification, the result of which is then combined with derived floor information to detect a fall. From a dataset of 15 people, we show that our fall detection system can achieve a high fall detection rate (97.08%) and a very low false detection rate (0.8%) in a simulated home environment.

Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/3293b84f599141d54a25bf6528bd2cf913de915c

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Excess-Risk of Distributed Stochastic Learners (2013, background): “function, say, EQ(w;x k;i). For example, in order to learn the hyper-plane that best separates feature data h k;i belonging to one of two classes y k(i) 2f+1; 1g, a support-vector-machine (SVM) [16], [17] would minimize the expected value of the following loss function over w (with the expectation c”
Survey on Fall Detection and Fall Prevention Using Wearable and External Sensors (2014, background): “In [9], a computer-vision fall detection system, image noise is eliminated by removing blobs with a size smaller than 50 pixels.”
Internet of Things: Architectures, Protocols, and Applications (2017, methodology): “[80] use computer vision based techniques for analyzing postures of the human body.”
A Survey on Activity Detection and Classification Using Wearable Sensors (2017, unknown): “Others are performing posture recognition [162].”
Deep learning for posture analysis in fall detection (2014, methodology): “…information related to the human being, we should subtract the environment part in the images acquired from the camera; to achieve this, the first thing in the method we choose is to build a codebook [6] and train the codebook for the background, then we can extract the silhouette of the person.”
Smart Homes for Elderly Healthcare—Recent Advances and Research Challenges (2017, background): “In case of non-injurious falls, around 47% of persons experiencing a fall need external support to get up [114].”