Search

One of the grand goals of robotics is to have assistive robots living side-by-side with humans, autonomously assisting humans in everyday activities. To be able to interact with humans and assist them, robots must be able to understand and interpret human activities. There is a growing interest in the problem of human activity recognition. Despite much progress, most computer vision researchers have narrowed the problem towards fixed camera viewpoint owing to inherent difficulty to train their systems across all possible viewpoints. However, since the robots and humans are free to move around in the environment, the viewpoint of a robot with respect to a person varies all the time. Therefore, we attempt to relax the infamous fixed viewpoint assumption and present a novel and efficient framework to recognize and classify human activities from monocular video source from arbitrary viewpoint. The proposed framework comprises of two stages: human pose recognition and human activity recognition. In the pose recognition stage, an ensemble of pose models performs inference on each video frame. Each pose model estimates the probability that the given frame contains the corresponding pose. Over a sequence of frames, each pose model forms a time series. In the activity recognition stage, we use nearest neighbor, with dynamic time warping as a distance measure, to classify pose time series. We have built a small-scale proof-of-concept model and performed some experiments on three publicly available datasets. The satisfactory experimental results demonstrate the efficacy of our framework and encourage us to further develop a full-scale architecture.

There is an increasing interest in real-time fall detection systems for the elderly in developed countries because more and more elderly are staying alone. There is a great demand for such fall detections systems in the smart home industry and the healthcare industry. Various fall detection approaches have been proposed recently by researchers. However, the majority of the proposed approaches require sensors to be attached on the subjects under surveillance. Sensors are intrusive and restrictive. Moreover, critical situations can often go undetected if the elderly forget to wear those vital sensors. As a result, researchers have recently gained interest in computer vision based solutions. Viewpoint invariance is a very important issue in computer vision because camera position is arbitrary and the subjects are free to move around in the environment. This paper presents a vision-based framework that can detect falls using a single camera, irrespective of the viewpoint of the camera with respect to the subjects. The proposed system makes use of invariant pose models which perform view-invariant human pose recognition. An ensemble of pose models performs inference on each video frame. Each pose model employs an expectation-maximization algorithm to estimate the probability that the given frame contains the corresponding pose. Over a sequence of frames, all the pose models collectively produce a multivariate time series. The system detects falls by analyzing the time series. We utilize the fuzzy hidden Markov model (FHMM) to detect falls. We have performed some experiments on two datasets and the results are found to be promising.

We present a fast, robust, fault-tolerant and nonparametric approach to reconstruct 3D human skeleton model from single images, regardless of the viewpoint of the camera relative to the observed person. The system achieves high degree of tolerance using search-based inference and probabilistic joint labeling. The system achieves 100% view invariance using a pruning-based parallel search algorithm.

Human activity recognition has become very popular in the field of computer vision. In this paper, we present a simple, robust and computationally efficient algorithm, architecture and implementation to recognise and classify human activities in real-time using very few training data. We employ a spatio-temporal representation of human activities by combining trajectory information and invariant spatial information of the subjects. Activities are classified by a support vector machine (SVM) with a radial basis kernel. Optimal parameters for the SVM are found through a 10-fold cross-validation. Experimental results demonstrate that the proposed system is effective and efficient. When tested on the Weizmann dataset, the system achieves a recognition rate above 90% for one-shot learning which is above benchmark scores in accordance with the literature. The system is also found to be robust against noise, deformation and variation in viewpoints. The system is feasible to operate efficiently in real-time and deployable in intelligent environments.