Combining image regions and human activity for indirect object recognition in indoor wide-angle views

Traditional methods of object recognition are reliant on shape and so are very difficult to apply in cluttered, wideangle and low-detail views such as surveillance scenes. To address this, a method of indirect object recognition is proposed, where human activity is used to infer both the location and identity of objects. No shape analysis is necessary. The concept is dubbed 'interaction signatures', since the premise is that a human will interact with objects in ways characteristic of the function of that object - for example, a person sits in a chair and drinks from a cup. The human-centred approach means that recognition is possible in low-detail views and is largely invariant to the shape of objects within the same functional class. This paper implements a Bayesian network for classifying region patches with object labels, building upon our previous work in automatically segmenting and recognising a human's interactions with the objects. Experiments show that interaction signatures can successfully find and label objects in low-detail views and are equally effective at recognising test objects that differ markedly in appearance from the training objects.

Explicit state duration HMM for abnormality detection in sequences of human activity

The importance of explicit duration modelling for classification of sequences of human activity and the reliable and timely detection of duration abnormality was highlighted. The normal classes of behavior were designed to highlight the importance of modelling duration given the limitations of the tracking system. It was found that HMM was the weakest model for classification of the unseen normal sequences with 81% accuracy. Long term abnormality was investigated by artificially varying the duration of primary activity in a randomly selected test sequence. The incorporation of duration in models of human behavior is an important consideration for systems seeking to provide cognitive support and to detect deviation in the behavorial patterns.

Human activity learning and segmentation using partially hidden discriminative models

Learning and understanding the typical patterns in the daily activities and routines of people from low-level sensory data is an important problem in many application domains such as building smart environments, or providing intelligent assistance. Traditional approaches to this problem typically rely on supervised learning and generative models such as the hidden Markov models and its extensions. While activity data can be readily acquired from pervasive sensors, e.g. in smart environments, providing manual labels to support supervised training is often extremely expensive. In this paper, we propose a new approach based on semi-supervised training of partially hidden discriminative models such as the conditional random field (CRF) and the maximum entropy Markov model (MEMM). We show that these models allow us to incorporate both labeled and unlabeled data for learning, and at the same time, provide us with the flexibility and accuracy of the discriminative framework. Our experimental results in the video surveillance domain illustrate that these models can perform better than their generative counterpart, the partially hidden Markov model, even when a substantial amount of labels are unavailable.

Learning feature trajectories using Gabor Filter Bank for human activity segmentation and recognition

We describe a novel method for human activity segmentation and interpretation in surveillance applications based on Gabor filter-bank features. A complex human activity is modeled as a sequence of elementary human actions like walking, running, jogging, boxing, hand-waving etc. Since human silhouette can be modeled by a set of rectangles, the elementary human actions can be modeled as a sequence of a set of rectangles with different orientations and scales. The activity segmentation is based on Gabor filter-bank features and normalized spectral clustering. The feature trajectories of an action category are learnt from training example videos using Dynamic Time Warping. The combined segmentation and the recognition processes are very efficient as both the algorithms share the same framework and Gabor features computed for the former can be used for the later. We have also proposed a simple shadow detection technique to extract good silhouette which is necessary for good accuracy of an action recognition technique. © 2008 IEEE.

A hybrid FMM-CART model for human activity recognition

In this paper, the application of a hybrid model combining the fuzzy min-max (FMM) neural network and the classification and regression tree (CART) to human activity recognition is presented. The hybrid FMM-CART model capitalizes the merits of both FMM and CART in data classification and rule extraction. To evaluate the effectiveness of FMM-CART, two data sets related to human activity recognition problems are conducted. The results obtained are higher than those reported in the literature. More importantly, practical rules in the form of a decision tree are extracted to provide explanation and justification for the predictions from FMM- CART. This outcome positively indicates the potential of FMM- CART in undertaking human activity recognition tasks.

Sand pads : a promising technique to quantify human visitation into nature conservation areas

In many places the expansion of urban areas has brought recreationists into close proximity to nature conservation areas, sometimes leading to conflict where recreation and sensitive environmental or natural values are incompatible. An important first step in managing these conflicts is to assess the degree and nature of the problem. We describe the application, and methodological considerations, associated with the use of an innovative, low-cost, practical technique to monitor human intrusions into a wetland reserve which has been designated as ‘off-limits’ to the general public. The use of seven frequently monitored sand pads over 13 weeks enabled us to determine that intrusions occurred in every week (3–14 per sand pad), deep inside the reserve during most weeks, and also identified the key access points. Most intrusions occurred during holiday periods and were by walkers or cyclists. We also conducted a series of simple experiments to examine the utility of sand pads. Our sand pads maintained their shape well and held footprints for over 1 month, they were rarely avoided by people and provided reliable indices of the level of human activity. Sand hardness varied with rainfall, and hardened sand was frequent (53.8% of 26 days) and potentially prevented detection of people. We conclude that the sand pad technique is an effective and efficient tool to measure recreational use of off-limits areas and other conservation and recreation areas, provided human traffic is not too intense, and that checks are made reasonably frequently.

Automated classification of human daily activities in ambulatory environment

This paper presents a human daily activity classification approach based on the sensory data collected from a single tri-axial accelerometer worn on waist belt. The classification algorithm was realized to distinguish 6 different activities including standing, jumping, sitting-down, walking, running and falling through three major steps: wavelet transformation, Principle Component Analysis (PCA)-based dimensionality reduction and followed by implementing a radial basis function (RBF) kernel Support Vector Machine (SVM) classifier. Two trials were conducted to evaluate different aspects of the classification scheme. In the first trial, the classifier was trained and evaluated by using a dataset of 420 samples collected from seven subjects by using a k-fold cross-validation method. The parameters σ and c of the RBF kernel were optimized through automatic searching in terms of yielding the highest recognition accuracy and robustness. In the second trial, the generation capability of the classifier was also validated by using the dataset collected from six new subjects. The average classification rates of 95% and 93% are obtained in trials 1 and 2, respectively. The results in trial 2 show the system is also good at classifying activity signals of new subjects. It can be concluded that the collective effects of the usage of single accelerometer sensing, the setting of the accelerometer placement and efficient classifier would make this wearable sensing system more realistic and more comfortable to be implemented for long-term human activity monitoring and classification in ambulatory environment, therefore, more acceptable by users.

Anthropogenic change and human health

The chapter provides and introduction to the impacts human activity have had and continue to have on human health, especially in the context of urbanization and increasing population density.

Remote monitoring system enabling cloud technology upon smart phones and inertial sensors for human kinematics

Stroke is a common neurological condition which is becoming increasingly common as the population ages. This entails healthcare monitoring systems suitable for home use, with remote access for medical professionals and emergency responders. The mobile phone is becoming the easy access tool for self-evaluation of health, but it is hindered by inherent problems including computational power and storage capacity. This research proposes a novel cloud based architecture of a biomedical system for a wearable motion kinematic analysis system which mitigates the above mentioned deficiencies of mobile devices. The system contains three subsystems: 1. Bio Kin WMS for measuring the acceleration and rotation of movement 2. Bio Kin Mobi for Mobile phone based data gathering and visualization 3. Bio Kin Cloud for data intensive computations and storage. The system is implemented as a web system and an android based mobile application. The web system communicates with the mobile application using an encrypted data structure containing sensor data and identifiable headings. The raw data, according to identifiable headings, is stored in the Amazon Relational Database Service which is automatically backed up daily. The system was deployed and tested in Amazon Web Services.

Human residential status and habitat quality affect the likelihood but not the success of lapwing breeding in an urban matrix

Wildlife living in the suburbs faces the challenge of dealing with human presence and yard management (including the occurrence of pets) which vary at the scale of the house block. This study examined the influence of ecological factors (e.g. extent of grass and food availability) and anthropogenic factors (e.g. human activity and garden usage) on breeding site choice and reproductive success of the ground-nesting masked lapwing Vanellus miles on Phillip Island, Australia. Lapwings nested less frequently in residential properties (high levels of human usage) compared with vacant blocks and holiday houses. They were also more likely to breed on properties with high food availability and larger areas of grass. None of these variables influenced clutch size or the probability of eggs hatching, although larger clutches and higher hatching rates tended to be associated with more food. This study shows that, for an urban exploiting species, habitat quality is not homogenous at the scale of the house block, and that human activity is avoided by a species generally considered highly tolerant of people.

Social inclusion, social movements, and the characteristics of late modernity

In discussing the ideology of social inclusion, this paper demonstrates that the composition of community groups in a period of late modernity is worthy of consideration. Although it would appear, on the surface at least, that previously stable community institutions, such as family, organised religion, trade unions, occupation and residential stability, and so on, are being challenged by a broad rejection of the once powerful tool of tradition, society's attachment to a belief in the symbolic value of community remains strong. In an environment however, in which the interaction and interdependence of human activity is subject to continual re-evaluation as the current processes of industrialisation and globalisation unfold, the template of what constitutes 'community' may need to be re-defined. It is to this end that the present paper is concerned, in that it seeks to identify new community formations. Of particular interest, is the rise and reach of modern day 'social movements', and why, when analysing the subject from a macro-sociological perspective, they have come to assume such a pivotal role in occupying community spaces left vacant by the demise of traditional social institutions. The paper is exploratory in its focus, using relevant literature to posit some broad theoretical themes, with the aim of presenting such themes to encourage a shift in community debates away from traditional concerns about 'who' and 'how many', towards questions of why new community forms are emerging.

Towards solutions for a liveable future: progress, practice, performance, people: Proceedings of the 41st Annual Conference of the Architectural Science Association ANZAScA

There is hope! Since Al Gore disclosed the inconvenient truth, the climate has changed. The time of denial is over; the era of well-informed action and sound development is with us. Sustainability has now moved from the fringe, into the mainstream of politics, society, architecture and building practice. In this new context architectural science will contribute to two main tasks: prevent further damage to our environment, and respond to challenges invoked by climate change.
The built environment and human activity within it account for a large part of the problem. Architectural science and architectural practice are part of the solution. The ANZAScA 2007 conference focuses on the solutions architectural science has to offer toward a liveable future through the following generic themes:
progress – the evaluation and improvement of built facilities, new and existing, in terms of energy intensity, financial reward and environmental impact.
practice – the relationship between our cultural heritage, new facility design, retrofit design and its realisation
through construction.
performance – the connection between building operation targets, validation of performance, and user comfort and interaction in new and existing environments.
people – the effect of space on user behaviour, user responsibility and social wealth.
In response to this challenge, architectural science researchers including students, educators, and practitioners at ANZAScA 2007, present a broad range of research activity and concern within the built environment from global issues down to the specific actions of individuals.