The main trend of computer graphics is the creation of photorealistic images however, there is increasing interest in the simulation of artistic and illustrative techniques. This thesis investigates a profile based technique for automatically extracting artistic sketches from regular grid digital elevation models. The results resemble those drawn by skilled cartographers and artists.The use of cartographic line simplification algorithms, which are usually applied to complex two-dimensional lines such as coastlines, allow a set of most important points on the terrain surface to be identified, these form the basis for sketching.This thesis also contains a wide ranging review of terrain representation techniques and suggests a new taxonomy.

This thesis is concerned with the problem of recognizing industrial objects rapidly and flexibly. The system design is based on a general strategy that consists of a generalized local feature detector, an extended learning algorithm and the use of unique structure of the objects. Thus, the system is not designed to be limited to the industrial environment. The generalized local feature detector uses the gradient image of the scene to provide a feature description that is insensitive to a range of imaging conditions such as object position, and overall light intensity. The feature detector is based on a representative point algorithm which is able to reduce the data content of the image without restricting the allowed object geometry. Thus, a major advantage of the local feature detector is its ability to describe and represent complex object structure. The reliance on local features also allows the system to recognize partially visible objects. The task of the learning algorithm is to observe the feature description generated by the feature detector in order to select features that are reliable over the range of imaging conditions of interest. Once a set of reliable features is found for each object, the system finds unique relational structure which is later used to recognize the objects. Unique structure is a set of descriptions of unique subparts of the objects of interest. The present implementation is limited to the use of unique local structure. The recognition routine uses these unique descriptions to recognize objects in new images. An important feature of this strategy is the transference of a large amount of processing required for graph matching from the recognition stage to the learning stage, which allows the recognition routine to execute rapidly. The test results show that the system is able to function with a significant level of insensitivity to operating conditions; The system shows insensitivity to its 3 main assumptions -constant scale, constant lighting, and 2D images- displaying a degree of graceful degradation when the operating conditions degrade. For example, for one set of test objects, the recognition threshold was reached when the absolute light level was reduced by 70%-80%, or the object scale was reduced by 30%-40%, or the object was tilted away from the learned 2D plane by 300-400. This demonstrates a very important feature of the learning strategy: It shows that the generalizations made by the system are not only valid within the domain of the sampled set of images, but extend outside this domain. The test results also show that the recognition routine is able to execute rapidly, requiring 10ms-500ms (on a PDP11/24 minicomputer) in the special case when ideal operating conditions are guaranteed. (Note: This does not include pre-processing time). This thesis describes the strategy, the architecture and the implementation of the vision system in detail, and gives detailed test results. A proposal for extending the system to scale independent 3D object recognition is also given.

Evidence is presented that in the early stages of design or composition the mental processes used by artists for visual invention require a different type of support from those used for visualising a nearly complete object. Most research into machine visualisation has as its goal the production of realistic images which simulate the light pattern presented to the retina by real objects. In contrast sketch attributes preserve the results of cognitive processing which can be used interactively to amplify visual thought. The traditional attributes of sketches include many types of indeterminacy which may reflect the artist's need to be "vague". Drawing on contemporary theories of visual cognition and neuroscience this study discusses in detail the evidence for the following functions which are better served by rough sketches than by the very realistic imagery favoured in machine visualising systems. 1. Sketches are intermediate representational types which facilitate the mental translation between descriptive and depictive modes of representing visual thought. 2. Sketch attributes exploit automatic processes of perceptual retrieval and object recognition to improve the availability of tacit knowledge for visual invention. 3. Sketches are percept-image hybrids. The incomplete physical attributes of sketches elicit and stabilise a stream of super-imposed mental images which amplify inventive thought. 4. By segregating and isolating meaningful components of visual experience, sketches may assist the user to attend selectively to a limited part of a visual task, freeing otherwise over-loaded cognitive resources for visual thought. 5. Sequences of sketches and sketching acts support the short term episodic memory for cognitive actions. This assists creativity, providing voluntary control over highly practised mental processes which can otherwise become stereotyped. An attempt is made to unite the five hypothetical functions. Drawing on the Baddeley and Hitch model of working memory, it is speculated that the five functions may be related to a limited capacity monitoring mechanism which makes tacit visual knowledge explicitly available for conscious control and manipulation. It is suggested that the resources available to the human brain for imagining nonexistent objects are a cultural adaptation of visual mechanisms which evolved in early hominids for responding to confusing or incomplete stimuli from immediately present objects and events. Sketches are cultural inventions which artificially mimic aspects of such stimuli in order to capture these shared resources for the different purpose of imagining objects which do not yet exist. Finally the implications of the theory for the design of improved machine systems is discussed. The untidy attributes of traditional sketches are revealed to include cultural inventions which serve subtle cognitive functions. However traditional media have many short-comings which it should be possible to correct with new technology. Existing machine systems for sketching tend to imitate nonselectively the media bound properties of sketches without regard to the functions they serve. This may prove to be a mistake. It is concluded that new system designs are needed in which meaningfully structured data and specialised imagery amplify without interference or replacement the impressive but limited creative resources of the visual brain.

Early image understanding seeks to derive analytic representations from image data. This thesis presents steps towards this goal for three-dimensional imagery by focusing on the inference of trace points (points belonging to surfaces), and the estimation of associated differential structure given by the principal curvature and direction fields over smooth surfaces. Computation of these fields is posed as the determination of a cross section through the bundle of curvature frames over the estimated trace. Algorithm robustness and the stability of results are essential for analysis of real images; to this end, I present a functional minimization algorithm where the principal direction cross section meets appropriate criteria for a minimum, and develop an implementation as an iterative constraint satisfaction procedure based on local surface smoothness properties. For shape description and eventual object recognition, the exact recovery of local structure everywhere is less important than the identification of singular surface points which prove stable to noise and small surface perturbations, in particular, the umbilic points of surfaces. Such points are computed naturally from the estimated local surface structure embodied in the principal direction cross section of the frame bundle. Examples of the recovery of local structure are presented for synthetic images degraded by noise and for clinical magnetic resonance images.

The model training algorithm is a critical component in the statistical pattern recognition approaches which are based on the Bayes decision theory. Conventional applications of the Bayes decision theory usually assume uniform error cost and result in a ubiquitous use of the maximum a posteriori (MAP) decision policy and the paradigm of distribution estimation as practice in the design of a statistical pattern recognition system. The minimum classification error (MCE) training method is proposed to overcome some substantial limitations for the conventional distribution estimation methods. In this thesis, three aspects of the MCE method are generalized. First, an optimal classifier/recognizer design framework is constructed, aiming at minimizing non-uniform error cost.A generalized training criterion named weighted MCE is proposed for pattern and speech recognition tasks with non-uniform error cost. Second, the MCE method for speech recognition tasks requires appropriate management of multiple recognition hypotheses for each data segment. A modified version of the MCE method with a new approach to selecting and organizing recognition hypotheses is proposed for continuous phoneme recognition. Third, the minimum verification error (MVE) method for detection-based automatic speech recognition (ASR) is studied. The MVE method can be viewed as a special version of the MCE method which aims at minimizing detection/verification errors. We present many experiments on pattern recognition and speech recognition tasks to justify the effectiveness of our generalizations.

Thesis (M.E.)--University of Canterbury, 2007. / Typescript (photocopy). "August 2007." Includes bibliographical references (leaves [105]-109). Also available via the World Wide Web.

Thesis (Ph.D.)--University of Wollongong, 2005. / Typescript. Includes bibliographical references: p. 229-243.

Thesis (Ph. D.)--State University of New York at Binghamton, Department of Systems Science and Industrial Engineering, Thomas J. Watson School of Engineering and Applied Science, 2007. / Includes bibliographical references.

Thesis (M.S.)--University of North Texas, Dec., 2007. / Title from title page display. Includes bibliographical references.

Thesis (M.Eng.Sc.)--University of Western Australia, 2004.