Behavior monitoring for assistive environments using multiple views

Abstract

This work presents an approach to behavior understanding using multiple cameras. This approach is appropriate for monitoring people in an assistive enivronment for the purpose of issuing alerts in cases of abnormal behavior. The output of multiple classifiers is used to model and extract abnormal behaviour from both the target trajectory and the target short term activity (i.e., walking, running, abrupt motion, etc). Spatial information is obtained after an offline camera registration using homography information. The proposed approach is verified experimentally in an indoor environment. The experiments are performed with a single moving target, however the method can be generalised to multiple moving targets, which may occlude each other, due to the use of multiple cameras.

Fusion of Color and Depth Video for Human Behavior Recognition in an Assistive Environment

“…There are several works on human behavior recognition in assistive environments using color cameras, e.g., [4], [5]. The color information captured by conventional cameras is a very useful cue, which can be used for environment modeling and object tracking.…”

Using optical flow and spectral clustering for behavior recognition and detection of anomalous behaviors

“…This method computes dense optical ow elds and Based on histograms of optical ow, proposes methods to automatically detect dangerous motion behavior in crowds exist. In [6] an approach to behavior understanding using multiple cameras is presented. From each camera some low level features from the foreground objects are extracted using optical ow calculation.…”

Human centered computing for the development of assistive environments

“…A popular method for action modeling is the utilization of Hidden Markov Models (HMM). In [5] the input from multiple cameras is fused to classify visual activities. However, the HMM requires high amount of training data, which means high effort for annotation and training.…”