Thesis (M. Phil.)--University of Hong Kong, 2002. / Includes bibliographical references (leaves 107-109).

Thesis--Ph. D., University of Hong Kong, 1986.

Thesis (M. Sc.)--York University, 2002. Graduate Programme in Computer Science. / Typescript. Includes bibliographical references (leaves 105-119). Also available on the Internet. MODE OF ACCESS via web browser by entering the following URL: http://wwwlib.umi.com/cr/yorku/fullcit?pMQ71638.

Includes bibliographical references (p. 59-66).

Traditional geometric scene analysis cannot attempt to address the understanding of human vision. Instead it adopts an algorithmic approach, concentrating on geometric model fitting. Human vision, however, is both quick and accurate but very little is known about how the recognition of objects is performed with such speed and efficiency. It is thought that there must be some process both for coding and storage which can account for these characteristics. In this thesis a more strict emulation of human vision, based on work derived from medical psychology and other fields, is proposed. Human beings must store perceptual information from which to make comparisons, derive structures and classify objects. It is widely thought by cognitive psychologists that some form of symbolic representation is inherent in this storage. Here a mathematical syntax is defined to perform this kind of symbolic description. The symbolic structures must be capable of manipulation and a set of operators is defined for this purpose. The early visual cortex and geniculate body are both inherently parallel in operation and simple in structure. A broadly connectionist emulation of this kind of structure is described, using independent computing elements, which can perform segmentation, re-colouring and generation of the base elements of the description syntax. Primal colour information is then collected by a second network which forms the visual topology, colouring and position information of areas in the image as well as a full description of the scene in terms of a more complex symbolic set. The idea of different visual contexts is introduced and a model is proposed for the accumulation of context rules. This model is then applied to a database of natural images.

Not available / text

Here we describe a gray scale object recognition system based on foveated corner finding, the computation of sequential fixation points, and elements of Lowe’s SIFT transform. The system achieves rotational, transformational, and limited scale invariant object recognition that produces recognition decisions using data extracted from sequential fixation points. It is broken into two logical steps. The first is to develop principles of foveated visual search and automated fixation selection to accomplish corner search. The result is a new algorithm for finding corners which is also a corner-based algorithm for aiming computed foveated visual fixations. In the algorithm, long saccades move the fovea to previously unexplored areas of the image, while short saccades improve the accuracy of putative corner locations. The system is tested on two natural scenes. As an interesting comparison study we compare fixations generated by the algorithm with those of subjects viewing the same images, whose eye movements are being recorded by an eyetracker. The comparison of fixation patterns is made using an information-theoretic measure. Results show that the algorithm is a good locator of corners, but does not correlate particularly well with human visual fixations. The second step is to use the corners located, which meet certain goodness criteria, as keypoints in a modified version of the SIFT algorithm. Two scales are implemented. This implementation creates a database of SIFT features of known objects. To recognize an unknown object, a corner is located and a feature vector created. The feature vector is compared with those in the database of known objects. The process is continued for each corner in the unknown object until enough information has been accumulated to reach a decision. The system was tested on 78 gray scale objects, hand tools and airplanes, and shown to perform well. / text

This work concerns the development of efficient methods for the recognition of binary picture objects, with a view to applications such as automatic location or inspection of industrial components and optical character recognition. Most common pattern recognition schemes attempt to overcome the high dimensionality of an input picture by taking feature measurements from it, which it is hoped will retain sufficient 'useful' information to enable correct classification to be made. If features are chosen on an 'ad hoc' basis as is often the case, there may be a loss of 'useful' information in the transformation from picture to feature space. Thus the error rate of the subsequent classifier may be increased. Furthermore, there is no easy way of estimating the increase in error rate. One way of ensuring zero loss of information in the transformation to feature space is to use a reversible transformation, which by suitable coding removes much of the redundancy in the original picture. A number of possible picture coding methods are examined. Of these, skeleton coding is chosen as being most suitable. The technique adopted thus involves the reduction of a binary picture object to a skeleton. By means of topological analysis and limb measurements, a feature vector is then produced suitable for subsequent classification by a nearest neighbour classifier. In order to be of practical use, any pattern recognition scheme must be capable of efficient implementation on readily available hardware. Thus much consideration has been given to the implementation of the algorithms described. In addition a study has been made of various types of hardware for picture processing. During the course of the work, a number of useful hardware and software tools were developed. These are described in some detail, and include an interactive system for the development of picture processing algorithms •

This thesis describes an investigation of the possibility of using interactive progr~ing languages for the production of animated films of models of 3-D solids. Two such languages are described, CGALt0 and CGALtl. CGAL#0 is an interactive command interpretive language and CGAL#l is an interpretive procedural high level language. Both these languages were especially designed for the purposes of 3-D model animation and implemented by the author. A dynamic display priority hidden surface algorithm, designed by the author, is presented. This algorithm is similar to the Newell algorithm but utilizes a different, more efficient priority sort. At present the author has implemented a restricted version of this algorithm. The restrictions are that the polygonal facets are non-intersecting and do not overlap one another cyclically. With cyclically overlapping facets this restricted algorithm does not guarantee correct results. This restricted algorithm sorts the list of facets in priority order without ever having to split a facet. An extended form of this algorithm intersecting and cyclically overlapping developed but not yet implemented. dealing with facets has been The author algorithm that hidden surface problem. has also designed and implemented an utilizes the priority list produced by the algorithm in order to solve the hidden line

published_or_final_version / abstract / Computer Science / Master / Master of Philosophy