Tracking multiple articulated objects (such as a human body) and handling occlusion between them is a challenging problem in automated video analysis. This work proposes a new approach for accurately and steadily visual tracking people, which should function even if the system encounters occlusion in video sequences. In this approach, targets are represented with a Gaussian mixture, which are adapted to regions of the target automatically using an EM-model algorithm. Field speeds are defined for changed pixels in each frame based on the probability of their belonging to a particular person's blobs. Pixels are matched to the models using a fast numerical level set method. Since each target is tracked with its blob's information, the system is capable of handling partial or full occlusion during tracking. Experimental results on a number of challenging sequences that were collected in non-experimental environments demonstrate the effectiveness of the approach.

Since online database applications have become increasingly used in clinical systems, accessing to an online immunization record system needs to be addressed to keep people updated about their latest immunization status and help providers to recommend the next appropriate vaccine at any location and anytime. Sufficient Health Information Systems can bridge the gap between the clinical and technical knowledge and benefit healthcare system. In this study, the requirement of designing a database for an immunization record model was reviewed, and a model was designed; subsequently, a database application was developed, and the qualitative assessment was deployed to evaluate the quality of data and some of usability factors. Through this study, the researcher describes how the data model was designed based on the information gained from Canadian resources such as Public Health Agency of Canada, Centers for Disease Controls, and Canadian Immunization Guide- seventh edition; then, a database application was developed, and the qualitative evaluation was performed to understand healthcare providers’ expectation from the real system. / Graduate

The effect hand action representations have on language processing was investigated using eye-tracking techniques. Subjects were shown an image of a hand action and asked to hold the action in working memory while reading a sentence, which described an actor lifting, or using an object. The displayed hand actions were related to either a functional (using) or volumetric (lifting) interaction with an object that matched or did not match the object mentioned in the sentence. A neutral condition was also used which displayed a black circle instead of a hand action. No significant difference was found between any of the five working memory conditions for gaze duration, probability of word skipping, and several other dependent measures utilized in the study. A significant difference was found for gaze duration when the conditions were restricted. Shorter gaze duration was observed for the hand action congruent to both the context and the object mentioned in the sentence and longer gaze duration was observed for the hand action congruent to only the sentence context. Some possible explanations of the results are that subjects may not have encoded the hand actions as action representations, or that hand actions represented in working memory have no effect on sentence processing. / Graduate / 0633

No description available.

Bacterial contamination due to excessive levels of bacteria is a confounding problem and remediation of impaired watersheds relies on the detection of fecal indicator bacteria and then assessing the source of said bacteria. Bacterial source tracking (BST) is an approach for assessing potential sources of this contamination. The purpose of this study was to utilize both cultivation-independent and –dependent methods to improve the ability to track sources of fecal contamination. First, E. coli community composition was assessed across three standard water quality assessments including USEPA Methods 1603 and 1604, and Colilert®, to determine their impact on BST library-based performance. Results indicate that the three assessed methods of enumeration and isolation may select for different populations of E. coli and standardized methods may be warranted if library-dependent BST is part of a research plan. Next, BST techniques were used to enumerate and characterize E. coli communities across various dairy manure management techniques used in the Leon River watershed in central Texas to determine effectiveness of BST efforts in tracking contamination from dairy manure. Results of this study indicated that manure and effluent management strategies which employed means to remove solids from the manure tended to decrease the levels of E. coli in the effluent. Some E. coli genotypes were found across the managerial treatments even though there were no clear seasonal trends or site groupings among the dataset. The vast majority of the isolates classified using the Texas E. coli BST library were correctly classified back to their major source class, thus increasing confidence in the methods currently being utilized to track dairy fecal contributions in this Central Texas watershed. Finally, deer bacterial fecal communities from south and central Texas were analyzed using 454-pyrosequencing to assess the potential for the development of a deer-specific BST marker. Microbial communities did not cluster by site or year suggesting that deer fecal communities in these Texas regions are stable over time and could be amenable to marker development.

A pose estimation method is proposed for measuring the position and orientation of a biopsy needle. The technique is to be used as a touchless needle guide system for guidance of percutaneous procedures with 4D ultrasound. A pair of uncalibrated, light-weight USB cameras are used as inputs. A database is prepared offline, using both the needle line estimated from camera-captured images and the true needle line recorded from an independent tracking device. A nonparametric learning algorithm determines the best fit model from the database. This model can then be used in real-time to estimate the true position of the needle with inputs from only the camera images. Simulation results confirm the feasibility of the method and show how a small, accurately made database can provide satisfactory results. In a series of tests with cameras, we achieved an average error of 2.4mm in position and 2.61° in orientation. The system is also extended to real ultrasound imaging, as the two miniature cameras capture images of the needle in air and the ultrasound system captures a volume as the needle moves through the workspace. A new database is created with the estimated 3D position of the needle from the ultrasound volume and the 2D position and orientation of the needle calculated from the camera images. This study achieved an average error of 0.94 mm in position and 3.93° in orientation.

Thesis (PhD) -- University of South Australia, 1999

Bibliography: leaves 106-110. / xii, 110 leaves : ill. ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Presents a novel system for automated track association, which provides significant improvement on previously proposed methods. The track association system is designed on the premise that the characteristic pattenrs of multimode propagation are a function of the ionospheric conditions. One component of the track association system models the multimode patterns over the radar coverage. An estimated multimode patterns can be compared to the patterns formed by observed tracks, to determine if the observed tracks are multimode tracks from a common target. The comparison of the estimated multimode patterns to be observed track patterns is performed with an association metric. The association metric requires certain parameters of the multimode patterns to be modelled over the radar coverage. / Thesis (Ph.D.)--University of Adelaide, Dept. of Electrical and Electronic Engineering, 1999

Title page, abstract and table of contents only. The complete thesis in print form is available from the University of Adelaide Library. / This thesis is concerned with the core computer vision challenge of obtaining efficient and robust visual tracking of objects over extended image sequences. Effective solutions to this problem are crucial for applications such as smart video surveillance, intelligent human machine interaction, and robotics. Most tracking algorithms can be classified into two major types, namely, probabilistic filtering algorithms and deterministic localisation algorithms. This thesis presents novel enhancements to both types of algorithm. The probabilistic filtering algorithms adopted in visual tracking are mainly based on Kalman filters and particle filters. Whereas Kalman filters are restricted to linear and Gaussian noise models, particle filters can propagate more general distributions, albeit only approximately. This is valuable in visual tracking, as simple models of noise do not suffice. Although particle filter trackers have been quite successful, they too have significant drawbacks. Several strategies are advanced in this thesis to overcome these limitations. Two alternative means are proposed for generating a proposal distribution, which is a key step in particle filtering. These increase the efficiency and robustness of the algorithm in the presence of sudden motion. The particle filter is also extended so as to accommodate multiple cues, such as colour and edge information, affording greater reliability. Additionally, an efficient kernel subspace method is introduced to capture a tracked object's appearance. Finally, a novel method is proposed for tracking the motion of an articulated structure. This significantly improves sampling efficiency, alleviating the curse of dimensionality in Monte Carlo sampling methods. The value of these enhancements is confirmed experimentally. The second part of this thesis concerns the mean shift algorithm, recently advanced as an alternative to stochastic trackers, that seeks the global mode of a suitable density function. A novel, multi-bandwidth mean shift procedure is presented along with a means of accelerating the algorithm. This improved tracker is applied to the problems of object localisation and visual tracking. We empirically show on various data sets that the proposed algorithm reliably finds the true object location when the initial position of the mean shift is far away from the global maximum, in contrast with the conventional mean shift algorithm which can often become trapped in a local maximum. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1229506 / Thesis (Ph.D.) -- University of Adelaide, School of Computer Science, 2006

Bibliography: leaves 106-110. / xii, 110 leaves : ill. ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Presents a novel system for automated track association, which provides significant improvement on previously proposed methods. The track association system is designed on the premise that the characteristic pattenrs of multimode propagation are a function of the ionospheric conditions. One component of the track association system models the multimode patterns over the radar coverage. An estimated multimode patterns can be compared to the patterns formed by observed tracks, to determine if the observed tracks are multimode tracks from a common target. The comparison of the estimated multimode patterns to be observed track patterns is performed with an association metric. The association metric requires certain parameters of the multimode patterns to be modelled over the radar coverage. / Thesis (Ph.D.)--University of Adelaide, Dept. of Electrical and Electronic Engineering, 1999