Efficient real-time face detection for high resolution surveillance applications

This paper presents an efficient face detection method suitable for real-time surveillance applications. Improved efficiency is achieved by constraining the search window of an AdaBoost face detector to pre-selected regions. Firstly, the proposed method takes a sparse grid of sample pixels from the image to reduce whole image scan time. A fusion of foreground segmentation and skin colour segmentation is then used to select candidate face regions. Finally, a classifier-based face detector is applied only to selected regions to verify the presence of a face (the Viola-Jones detector is used in this paper). The proposed system is evaluated using 640 x 480 pixels test images and compared with other relevant methods. Experimental results show that the proposed method reduces the detection time to 42 ms, where the Viola-Jones detector alone requires 565 ms (on a desktop processor). This improvement makes the face detector suitable for real-time applications. Furthermore, the proposed method requires 50% of the computation time of the best competing method, while reducing the false positive rate by 3.2% and maintaining the same hit rate.

MAVwork : a framework for unified interfacing between micro aerial vehicles and visual controllers

Debugging control software for Micro Aerial Vehicles (MAV) can be risky out of the simulator, especially with professional drones that might harm people around or result in a high bill after a crash. We have designed a framework that enables a software application to communicate with multiple MAVs from a single unified interface. In this way, visual controllers can be first tested on a low-cost harmless MAV and, after safety is guaranteed, they can be moved to the production MAV at no additional cost. The framework is based on a distributed architecture over a network. This allows multiple configurations, like drone swarms or parallel processing of drones' video streams. Live tests have been performed and the results show comparatively low additional communication delays, while adding new functionalities and flexibility. This implementation is open-source and can be downloaded from github.com/uavster/mavwork

On the statistical determination of optimal camera configurations in large scale surveillance networks

The selection of optimal camera configurations (camera locations, orientations etc.) for multi-camera networks remains an unsolved problem. Previous approaches largely focus on proposing various objective functions to achieve different tasks. Most of them, however, do not generalize well to large scale networks. To tackle this, we introduce a statistical formulation of the optimal selection of camera configurations as well as propose a Trans-Dimensional Simulated Annealing (TDSA) algorithm to effectively solve the problem. We compare our approach with a state-of-the-art method based on Binary Integer Programming (BIP) and show that our approach offers similar performance on small scale problems. However, we also demonstrate the capability of our approach in dealing with large scale problems and show that our approach produces better results than 2 alternative heuristics designed to deal with the scalability issue of BIP.

SAIVT-QUT@TRECVid 2012 : interactive surveillance event detection

In this paper, we propose an approach which attempts to solve the problem of surveillance event detection, assuming that we know the definition of the events. To facilitate the discussion, we first define two concepts. The event of interest refers to the event that the user requests the system to detect; and the background activities are any other events in the video corpus. This is an unsolved problem due to many factors as listed below: 1) Occlusions and clustering: The surveillance scenes which are of significant interest at locations such as airports, railway stations, shopping centers are often crowded, where occlusions and clustering of people are frequently encountered. This significantly affects the feature extraction step, and for instance, trajectories generated by object tracking algorithms are usually not robust under such a situation. 2) The requirement for real time detection: The system should process the video fast enough in both of the feature extraction and the detection step to facilitate real time operation. 3) Massive size of the training data set: Suppose there is an event that lasts for 1 minute in a video with a frame rate of 25fps, the number of frames for this events is 60X25 = 1500. If we want to have a training data set with many positive instances of the event, the video is likely to be very large in size (i.e. hundreds of thousands of frames or more). How to handle such a large data set is a problem frequently encountered in this application. 4) Difficulty in separating the event of interest from background activities: The events of interest often co-exist with a set of background activities. Temporal groundtruth typically very ambiguous, as it does not distinguish the event of interest from a wide range of co-existing background activities. However, it is not practical to annotate the locations of the events in large amounts of video data. This problem becomes more serious in the detection of multi-agent interactions, since the location of these events can often not be constrained to within a bounding box. 5) Challenges in determining the temporal boundaries of the events: An event can occur at any arbitrary time with an arbitrary duration. The temporal segmentation of events is difficult and ambiguous, and also affected by other factors such as occlusions.

A tool for internet chatroom surveillance

Internet chatrooms are common means of interaction and communications, and they carry valuable information about formal or ad-hoc formation of groups with diverse objectives. This work presents a fully automated surveillance system for data collection and analysis in Internet chatrooms. The system has two components: First, it has an eavesdropping tool which collects statistics on individual (chatter) and chatroom behavior. This data can be used to profile a chatroom and its chatters. Second, it has a computational discovery algorithm based on Singular Value Decomposition (SVD) to locate hidden communities and communication patterns within a chatroom. The eavesdropping tool is used for fine tuning the SVD-based discovery algorithm which can be deployed in real-time and requires no semantic information processing. The evaluation of the system on real data shows that (i) statistical properties of different chatrooms vary significantly, thus profiling is possible, (ii) SVD-based algorithm has up to 70-80% accuracy to discover groups of chatters.

Multi-device key management using visual side channels in pervasive computing environments

In the modern connected world, pervasive computing has become reality. Thanks to the ubiquity of mobile computing devices and emerging cloud-based services, the users permanently stay connected to their data. This introduces a slew of new security challenges, including the problem of multi-device key management and single-sign-on architectures. One solution to this problem is the utilization of secure side-channels for authentication, including the visual channel as vicinity proof. However, existing approaches often assume confidentiality of the visual channel, or provide only insufficient means of mitigating a man-in-the-middle attack. In this work, we introduce QR-Auth, a two-step, 2D barcode based authentication scheme for mobile devices which aims specifically at key management and key sharing across devices in a pervasive environment. It requires minimal user interaction and therefore provides better usability than most existing schemes, without compromising its security. We show how our approach fits in existing authorization delegation and one-time-password generation schemes, and that it is resilient to man-in-the-middle attacks.

The feral horse foot. Part B : radiographic, gross visual and histopathological parameters of foot health in 100 Australian feral horses

Background It has been proposed that the feral horse foot is a benchmark model for foot health in horses. However, the foot health of feral horses has not been formally investigated. Objectives To investigate the foot health of Australian feral horses and determine if foot health is affected by environmental factors, such as substrate properties and distance travelled. Methods Twenty adult feral horses from five populations (n = 100) were investigated. Populations were selected on the basis of substrate hardness and the amount of travel typical for the population. Feet were radiographed and photographed, and digital images were surveyed by two experienced assessors blinded to each other's assessment and to the population origin. Lamellar samples from 15 feet from three populations were investigated histologically for evidence of laminitis. Results There was a total of 377 gross foot abnormalities identified in 100 left forefeet. There were no abnormalities detected in three of the feet surveyed. Each population had a comparable prevalence of foot abnormalities, although the type and severity of abnormality varied among populations. Of the three populations surveyed by histopathology, the prevalence of chronic laminitis ranged between 40% and 93%. Conclusions Foot health appeared to be affected by the environment inhabited by the horses. The observed chronic laminitis may be attributable to either nutritional or traumatic causes. Given the overwhelming evidence of suboptimal foot health, it may not be appropriate for the feral horse foot to be the benchmark model for equine foot health.

Image based visual servo control for fixed wing UAVs tracking linear infrastructure in wind

This paper presents an Image Based Visual Servo control design for Fixed Wing Unmanned Aerial Vehicles tracking locally linear infrastructure in the presence of wind using a body fixed imaging sensor. Visual servoing offers improved data collection by posing the tracking task as one of controlling a feature as viewed by the inspection sensor, although is complicated by the introduction of wind as aircraft heading and course angle no longer align. In this work it is shown that the effects of wind alter the desired line angle required for continuous tracking to equal the wind correction angle as would be calculated to set a desired course. A control solution is then sort by linearizing the interaction matrix about the new feature pose such that kinematics of the feature can be augmented with the lateral dynamics of the aircraft, from which a state feedback control design is developed. Simulation results are presented comparing no compensation, integral control and the proposed controller using the wind correction angle, followed by an assessment of response to atmospheric disturbances in the form of turbulence and wind gusts

Featureless visual processing for SLAM in changing outdoor environments

Vision-based SLAM is mostly a solved problem providing clear, sharp images can be obtained. However, in outdoor environments a number of factors such as rough terrain, high speeds and hardware limitations can result in these conditions not being met. High speed transit on rough terrain can lead to image blur and under/over exposure, problems that cannot easily be dealt with using low cost hardware. Furthermore, recently there has been a growth in interest in lifelong autonomy for robots, which brings with it the challenge in outdoor environments of dealing with a moving sun and lack of constant artificial lighting. In this paper, we present a lightweight approach to visual localization and visual odometry that addresses the challenges posed by perceptual change and low cost cameras. The approach combines low resolution imagery with the SLAM algorithm, RatSLAM. We test the system using a cheap consumer camera mounted on a small vehicle in a mixed urban and vegetated environment, at times ranging from dawn to dusk and in conditions ranging from sunny weather to rain. We first show that the system is able to provide reliable mapping and recall over the course of the day and incrementally incorporate new visual scenes from different times into an existing map. We then restrict the system to only learning visual scenes at one time of day, and show that the system is still able to localize and map at other times of day. The results demonstrate the viability of the approach in situations where image quality is poor and environmental or hardware factors preclude the use of visual features.

Supporting the development of inclusive products : the effects of everyday ambient illumination levels and contrast on older adults' near visual acuity

Current older adult capability data-sets fail to account for the effects of everyday environmental conditions on capability. This article details a study that investigates the effects of everyday ambient illumination conditions (overcast, 6000 lx; in-house lighting, 150 lx and street lighting, 7.5 lx) and contrast (90%, 70%, 50% and 30%) on the near visual acuity (VA) of older adults (n= 38, 65-87 years). VA was measured at a 1-m viewing distance using logarithm of minimum angle of resolution (LogMAR) acuity charts. Results from the study showed that for all contrast levels tested, VA decreased by 0.2 log units between the overcast and street lighting conditions. On average, in overcast conditions, participants could detect detail around 1.6 times smaller on the LogMAR charts compared with street lighting. VA also significantly decreased when contrast was reduced from 70% to 50%, and from 50% to 30% in each of the ambient illumination conditions. Practitioner summary: This article presents an experimental study that investigates the impact of everyday ambient illumination levels and contrast on older adults' VA. Results show that both factors have a significant effect on their VA. Findings suggest that environmental conditions need to be accounted for in older adult capability data-sets/designs.

Improving the measurement and surveillance of child abuse in Queensland emergency departments

There are no population studies of prevalence or incidence of child maltreatment in Australia. Child protection data gives some understanding but is restricted by system capacity and definitional issues across jurisdictions. Child protection data currently suggests that numbers of reports are increasing yearly, and the child protection system then becomes focussed on investigating all reports and diluting available resources for those children who are most in need of intervention. A public health response across multiple agencies enables responses to child safety across the entire population. All families are targeted at the primary level; examples include ensuring all parents know the dangers of shaking a baby or teaching children to say no if a situation makes them uncomfortable. The secondary level of prevention targets families with a number of risk factors, for example subsidised child care so children aren't left unsupervised after school when both parents have to be at work or home visiting for drug-addicted parents to ensure children are cared for. The tertiary response then becomes the responsibility of the child protection system and is reserved for those children where abuse and neglect are identified. This model requires that child safety is seen in a broader context than just the child protection system, and increasingly health professionals are being identified as an important component in the public health framework. If all injury is viewed as preventable and considered along a continuum of 'accidental' through to 'inflicted', it becomes possible to conceptualise child maltreatment in an injury context. Parental intent may not be to cause harm to the child, but by lack of insight or concern about risk, the potential for injury is high. The mechanisms for unintentional and intentional injury overlap and some suggest that by segregating child abuse (with the possible exception of sexual abuse) from unintentional injury, child abuse is excluded from the broader injury prevention initiative that is gaining momentum in the community. This research uses a public health perspective, specifically that of injury prevention, to consider the problem of child abuse. This study employed a mixed method design that incorporates secondary data analysis, data linkage and structured interviews of different professional groups. Datasets from the Queensland Injury Surveillance Unit (QISU) and The Department of Child Safety (DCS) were evaluated. Coded injury data was grouped according to intent of injury according to those with a code that indicated the ED presentation was due to child abuse, a code indicating that the injury was possibly due to abuse or, in the third group, the intent code indicated that the injury was unintentional and not due to abuse. Primary data collection from ED records was undertaken and information recoded to assess reliability and completeness. Emergency department data (QISU) was linked to Department of Child Safety Data to examine concordance and data quality. Factors influencing the collection and collation of these data were identified through structured interview methodology and analysed using qualitative methods. Secondary analysis of QISU data indicated that codes lacking specific information on the injury event were more likely to also have an intent code indicating abuse than those records where there was specific information on the injury event. Codes for abuse appeared in only 1.2% of the 84,765 records analysed. Unintentional injury was the most commonly coded intent (95.3%). In the group with a definite abuse code assigned at triage, 83% linked to a record with DCS and cases where documentation indicated police involvement were significantly more likely to be associated with a DCS record than those without such documentation. In those coded with an unintentional injury code, 22% linked to a DCS record with cases assigned an urgent triage category more likely to link than those with a triage category for resuscitation and children who presented to regional or remote hospitals more likely to link to a DCS record than those presenting to urban hospitals. Twenty-nine per cent of cases with a code indicating possible abuse linked to a DCS record. In documentation that indicated police involvement in the case, a code for unspecified activity when compared to cases with a code indicating involvement in a sporting activity and children less than 12 months of age compared to those in the 13-17 year old age group were all variables significantly associated with linkage to a DCS record. Only 13% of records contained documentation indicating that child abuse and neglect were considered in the diagnosis of the injury despite almost half of the sample having a code of abuse or possible abuse. Doctors and nurses were confident in their knowledge of the process of reporting child maltreatment but less confident about identifying child abuse and neglect and what should be reported. Many were concerned about implications of reporting, for the child and family and for themselves. A number were concerned about the implications of not reporting, mostly for the wellbeing of the child and a few in terms of their legal obligations as mandatory reporters. The outcomes of this research will help improve the knowledge of barriers to effective surveillance of child abuse in emergency departments. This will, in turn, ensure better identification and reporting practises; more reliable official statistical collections and the potential of flagging high-risk cases to ensure adequate departmental responses have been initiated.

Image-based visual navigation for mobile robots

We introduce a new image-based visual navigation algorithm that allows the Cartesian velocity of a robot to be defined with respect to a set of visually observed features corresponding to previously unseen and unmapped world points. The technique is well suited to mobile robot tasks such as moving along a road or flying over the ground. We describe the algorithm in general form and present detailed simulation results for an aerial robot scenario using a spherical camera and a wide angle perspective camera, and present experimental results for a mobile ground robot.

Implementation guide for surveillance of Clostridium difficile infection

The Implementation Guide for hospital surveillance of Clostridium difficile infection (CDI) has been produced by the Healthcare Associated Infection (HAI) Technical Working Group of the Australian Commission on Safety and Quality in Health Care (ACSQHC), and endorsed by the HAI Advisory Group. State jurisdictions and the ACSQHC have representatives on the Technical Working Group, and have had input into this document. The Guide is intended to be used by Australian hospitals and organisations to support the implementation of hospital-identified Clostridium difficile infection (CDI) surveillance using the endorsed case definition in this guide. It has been produced to support consistency of surveillance activities and is not intended to replace clinical assessment of infection for patient management.