No description available.

The recovery of generic solid parts is a fundamental step towards the realization of general-purpose vision systems. This thesis investigates issues in grouping, segmentation and recognition of parts from two-dimensional edge images. A new paradigm of part-based grouping of features is introduced that bridges the classical grouping and model-based approaches with the purpose of directly recovering parts from real images, and part-like models are used that both yield low theoretical complexity and reliably recover part-plausible groups of features. The part-like models used are statistical point distribution models, whose training set is built using random deformable superellipse. The computational approach that is proposed to perform model-guided part-based grouping consists of four distinct stages. In the first stage, codons, contour portions of similar curvature, are extracted from the raw edge image. They are considered to be indivisible image features because they have the desirable property of belonging either to single parts or joints. In the second stage, small seed groups (currently pairs, but further extension are proposed) of codons are found that give enough structural information for part hypotheses to be created. The third stage consists in initialising and pre-shaping the models to all the seed groups and then performing a full fitting to a large neighbourhood of the pre-shaped model. The concept of pre-shaping to a few significant features is a relatively new concept in deformable model fitting that has helped to dramatically increase robustness. The initialisations of the part models to the seed groups is performed by the first direct least-square ellipse fitting algorithm, which has been jointly discovered during this research; a full theoretical proof of the method is provided. The last stage pertains to the global filtering of all the hypotheses generated by the previous stages according to the Minimum Description Length criterion: the small number of grouping hypotheses that survive this filtering stage are the most economical representation of the image in terms of the part-like models. The filtering is performed by the maximisation of a boolean quadratic function by a genetic algorithm, which has resulted in the best trade-off between speed and robustness. Finally, images of parts can have a pronounced 3D structure, with ends or sides clearly visible. In order to recover this important information, the part-based grouping method is extended by employing parametrically deformable aspects models which, starting from the initial position provided by the previous stages, are fitted to the raw image by simulated annealing. These models are inspired by deformable superquadrics but are built by geometric construction, which render them two order of magnitudes faster to generate than in previous works. A large number of experiments is provided that validate the approach and, since several new issues have been opened by it, some future work is proposed.

Thesis (M. Ed.)--University of Hong Kong, 2006. / Title proper from title frame. Also available in printed format.

As more and more services are provided via WWW, how to reduce the perceived delay in WWW interaction becomes very import to the service providers to keep their users. Pre-fetching is an important technique for reducing latency in distributes systems like the WWW. Page pre-fetching takes advantage of local machine idle period of user viewing current page to deliver pages that user like to access in the near future. Being motivated by pre-fetching ideas and its practical bothers, we propose a server-initiated page pre-fetching method: XML enabled page-grouping to reduce Web latency. In our page-grouping scheme, we anticipate the page that user will like to access in the near future based on hyperlink and referral access probabilities of each page. The predictive pages are grouped and converted into a XML file embedding in the page that user currently requests. If the user clicks the predictive linked page, the corresponding HTML is regenerated directly from the embedded XML document. The proposed scheme allows alternative of batch grouping or on-line grouping. For the reason of avoiding server extra load, we suggest that the task of grouping static pages is performed periodically at server off-peak loading time. Beside static pages, we also hope to group dynamic page generated by CGI and illustrate the feasibility with an example of Web-based database query. When compared to previous page pre-fetching techniques, our page-grouping method has simplicity and practicability. By using XML document, add-on application modules are no more needed because that the XML processor is supported in new generator browsers like Microsoft IE 5.0. Furthermore, the way of converting grouping pages into embedded XML document makes predictive pages transparent to the proxy servers and the server side speculative service can work no matter whether there are proxy servers between the server and clients. Using trace simulations based on the logs of HTTP server http://www.kcg.gov.tw, we show that 67.84% URL request is referral request. It means that the probability is about 2/3 that users retrieve next Web page by clicking hyperlinks on currently viewing page. The logs are categorized according to the kind of official service. And the statistical results of every class of logs indicate that page always has a persistent referral access probabilities for a few days. It encourages us to get high hit rate of a predictive page by selecting it according to its high referral access probability. Considering bandwidth tradeoff, we discuss hit rate, traffic increase due to grouping and traffic intensity based on M/M/1 model. For online grouping of dynamic page, we take an example of database querying page on our simulating HTTP server. The experiment result leads to the conclusion that page grouping of pages of Web-based database querying can reduce server load of CGI processing, as the hit rate of the next page is about 18.48%.

Thesis (M.A.(Ed.))--University of Hong Kong, 1968. / Mimeographed.

Thesis (Ed. D.)--Boise State University, 2003. / Includes abstract. Includes bibliographical references (leaves 57-64). Also available online via the ProQuest Digital Dissertations database.

No abstract available. / Thesis (M.A.)-University of Durban-Westville, 1980.

No description available.

Thesis (Ph.D.)--University of Nebraska-Lincoln, 2008. / Title from title screen (site viewed Sept. 18, 2008). PDF text: ix, 175 p. ; 782 K. UMI publication number: AAT 3303507. Includes bibliographical references. Also available in microfilm and microfiche formats.

Thesis (M.A.(Ed.))--University of Hong Kong, 1968. / Also available in print.