No description available.

No description available.

No description available.

Thesis (M.S.)--Montana State University--Bozeman, 2004. / Typescript. Chairperson, Graduate Committee: Rockford Ross. Includes bibliographical references (leaves 37-38).

No description available.

The huge number of cctv cameras and security applications places increasing requirements on automatic visual tracking and behaviour classification systems. The best working example of such a tracker is the human visual system (hvs) which can flawlessly detect, track and understand almost any object or event. The research described in this thesis uses lessons learnt from studies of the hvs to develop a novel approach for computerbased visual tracking. In this approach, initial detection of moving objects is achieved using a new motion distillation paradigm which employs spatio-temporal wavelet decomposition of video. The method is shown to be more robust than traditional background modelling techniques while being computationally less expensive. As with the hvs, the approach uses a dual-channel tracking architecture to perform tracking. The motion channel, generated through motion distillation, handles object detection and initialises tracking. The form channel is used to resolve tracking ambiguities and occlusions. Qualitative and quantitative tracking results illustrate the advantages of this approach. This thesis also describes a new approach to the task of ob- 4 ject (e.G. Human) behaviour analysis - a subject which is of great importance, yet which is still an under-researched aspect of visual tracking. In the work described here, objects are categorised into vehicles, pedestrians, runners, groups and unknown pedestrian behaviour.

Djang et al. [1998] introduced a new model of types for declarative visual programming languages (VPLs). Implicit static typing is used in their type model, in order to eliminate the programming mechanisms associated with type declarations, provide immediate visual feedback with respect to type errors and guarantee type safe programs. Their type model also evaluates types on a level of granularity that is finer than in previous approaches to types. Instead of evaluating types on the basis of abstract names, their model determines a set of operations that an object guarantees and compares this set to the set of operations this object is required to support. If the set of required operations is a subset of the set of guaranteed operations, then the object is considered type safe. This granularity provides their model with the ability to support inheritance without introducing explicit type declarations and to communicate type errors to users without requiring the user to understand a large set of terminology. These features of their model attempt to provide VPL users with more powerful programming capabilities without the introduction of a high learning curve. In this thesis, an implementation of the Djang et al.'s model of types is presented. Data structures and algorithms are developed that conform to the axioms prescribed by Djang et al. The space and time complexity analyses for our data structures and algorithms are examined. Our implementation provides new insights into the cost and performance of the Djang et al.'s type model. / Graduation date: 2000

We believe concreteness, direct manipulation and responsiveness in a visual programming language increase its usefulness. However, these characteristics present a challenge in generalizing programs for reuse, especially when concrete examples are used as one way of achieving concreteness. In this thesis, we present a technique to solve this problem by deriving generality automatically through the analysis of logical relationships among concrete program entities from the perspective of a particular computational goal. Use of this technique allows a fully general form-based program with reusable abstractions to be derived from one that was specified in terms of concrete examples and direct manipulation. Also addressed in this thesis is how to statically represent the generalized programs. In general, we address how to design better static representations. A weakness of many interactive visual programming languages is their static representations. Lack of an adequate static representation places a heavy cognitive burden on a VPL's programmers, because they must remember potentially long dynamic sequences of screen displays in order to understand a previously-written program. However, although this problem is widely acknowledged, research on how to design better static representations for interactive VPLs is still in its infancy. Building upon the cognitive dimensions developed for programming languages by cognitive psychologists Green and others, we have developed a set of concrete benchmarks for VPL designers to use when designing new static representations. These benchmarks provide design-time information that can be used to improve a VPL's static representation. / Graduation date: 1997

Most current DFVPLs support flow control to facilitate experiments and complex problems. However, current approaches in DFVPLs still remain inefficient. We show that inadequacies in existing visual programming languages may be magnified in applications involving image analysis. These include a lack of efficient communication mechanisms and strong dependency on human involvement to customise properties. For instance, properties in one computational component can not be shared for other components. Moreover, conditional expressions used in control components hold data values that are unrelated with those computational components. Furthermore, since image processing libraries usua.lly only explicitly support pipeline processing, as exemplified by the widely used Insight Toolkit for Medical Image Segmentation and Registration (ITK), a looping algorithm would be difficult to implement without a feedback mechanism supported by the visual language itself. We propose a data-flow visual programming language that encompasses several novel control constructs and parameterised computational units. These components are facilitated by a novel hybrid data-flow model. We also present several conceptual models and design alternatives for control constructs. Several mechanisms and techniques are provided to enhance data propagation for these components. We demonstrate, in an environment that utilises ITK as the underlying processing engine, that the inadequacies in existing DFVPLs can be satisfactorily addressed through the visual components proposed in this thesis.

Thesis (M.S.)--Oregon State University, 2007. / Printout. Includes bibliographical references (leaves 37-39). Also available on the World Wide Web.