published_or_final_version / abstract / Electrical and Electronic Engineering / Master / Master of Philosophy

published_or_final_version / Electrical and Electronic Engineering / Master / Master of Philosophy

The information transfer efficiency of any quantum detection imaging system can be described by a unique measure: the detective quantum efficiency {DQE(f)}, which is a function of the statistically independent frequency channels. The DQE(f) is a combined descriptor which takes into account the signal transfer as well as noise transfer properties of a complete system. For a complicated multistage imaging system, each system component contributes noise. In this dissertation, physical and mathematical models for noise analysis are developed and verified experimentally with an x-ray image intensifier (XRII)-based imaging system. In such a system, the DQE at low frequency range is primarily determined by the x-ray detection and scintillation processes at the CsI layer of the XRII. The effects of x-ray photon energy and sensor layer thickness on DQE are measured in detail. Numerical calculations based on a physical model of x-ray interactions show a general agreement with the experimental data. At higher frequencies, the DQE behavior becomes more complicated. A mathematical model which combines the micro-image properties and noise statistics is formulated to analyze the noise power spectrum (NPS) of a linear n-stage imaging system. Measurement of NPS components of an XRII system verifies the validity of this analytical prediction. The associated image transfer properties are also measured with emphasis on the effect of signal-induced background on the image information transfer. The low frequency data derived from these image property measurements show further agreement with the numerical calculations based on the physical model. As a result of this predictability of information transfer efficiency, system gain and recording capacity are emphasized in the design consideration of a projected high performance XRII radiographic system.

No description available.

The design of parallel architectures to perform image segmentation processing is given. In addition, the various designs are evaluated as to their performance, and a discussion of an optimal design is given. In this thesis, a set of eight segmentation algorithms has been provided as a starting point. Four of these algorithms will be evaluated and partitioned using two techniques. From this study of partitioning and considering the data flow through the total system, architectures utilizing parallel techniques will be derived. Timing analysis using pen and paper techniques will be given on the architectures using three of today's current technologies. Next, NETWORK II.5 simulations will be run to provide performance measures. Finally, evaluations of the various architectures will be made as well as the applicability of using NETWORK II.5 as a simulation language.

<p>Much work has been done in building systems that can recognize gestures, e.g. as a component of sign language recognition systems. These systems typically use whole gestures as the smallest unit for recognition. Although high recognition rates have been reported, these systems do not scale well and are computationally intensive. The reason why these systems generally scale poorly is that they recognize gestures by building individual models for each separate gesture / as the number of gestures grows, so does the required number of models. Beyond a certain threshold number of gestures to be recognized, this approach become infeasible. This work proposed that similarly good recognition rates can be achieved by building models for subcomponents of whole gestures, so-called cheremes. Instead of building models for entire gestures, we build models for cheremes and recognize gestures as sequences of such cheremes. The assumption is that many gestures share cheremes and that the number of cheremes necessary to describe gestures is much smaller than the number of gestures. This small number of cheremes then makes it possible to recognized a large number of gestures with a small number of chereme models. This approach is akin to phoneme-based speech recognition systems where utterances are recognized as phonemes which in turn are combined into words.</p>

Durban is the busiest container port in Africa and there are plans for significant expansion during the next few years. This expansion includes the widening of the port entrance channel to accommodate larger "post-panamax" vessels. Complex crosscurrents near the port entrance, coupled with severe wind and wave conditions, may lead to the intermittent closure of the port which in turn could lead to significant economic implications. Information on the nature of the crosscurrents and how they affect the ships could assist harbour pilots in developing their skills and reduce the risks associated with steering ships into the port. The research involved a case study to develop an innovative new method for directly measuring the effect of wind, waves and crosscurrents on ships entering the port. The technology is based on the application of digital image processing to track the position of ships as they manoeuvre in the port approach channel. The key innovation of this research is the extraction of the heading direction of the ship from the image data. The angle between this heading direction and the true velocity vector (the "crab angle") is then a direct measure of the cross-track drift velocity (CTDV) due to the combined effects of wind, waves and currents. The crosscurrents are usually the main contributing factor to the cross-track drift. The aim of this research was to develop a fully automated image processing system for real-time ship monitoring, and to determine cross-track drift within a wide range of weather conditions and ship parameters. The methodology presented in this research allows the spatial structure of the CTDV along the harbour approach channel to be studied. The relationship of the CTDV to local surface winds was analysed. For deep draught vessels, measured CTDVs were found to be poorly correlated to surface winds. The spatial structure of the measured CTDVs shows distinct regions along the approach channel where vessels experience significantly larger drift velocities. In summary, with the software tools developed by this research, digital images can be captured automatically and analysed to produce ship tracks and crab angles. From this information an extensive database for ship manoeuvring in the approach channel can be developed and the safety and efficiency of port operations improved. / Thesis (M.Sc.Eng.)-University of Natal, Durban, 2002.

In this dissertation, first, we have proposed and implemented a new perceptually tuned wavelet based, rate scalable, and color image encoding/decoding system based on the human perceptual model. It is based on state-of-the-art research on embedded wavelet image compression technique, Contrast Sensitivity Function (CSF) for Human Visual System (HVS) and extends this scheme to handle optimal bit allocation among multiple bands, such as Y, Cb, and Cr. Our experimental image codec shows very exciting results in compression performance and visual quality comparing to the new wavelet based international still image compression standard - JPEG 2000. On the other hand, our codec also shows significant better speed performance and comparable visual quality in comparison to the best codec available in rate scalable color image compression - CSPIHT that is based on Set Partition In Hierarchical Tree (SPIHT) and Karhunen-Loeve Transform (KLT). Secondly, a novel wavelet based interframe compression scheme has been developed and put into practice. It is based on the Flexible Block Wavelet Transform (FBWT) that we have developed. FBWT based interframe compression is very efficient in both compression and speed performance. The compression performance of our video codec is compared with H263+. At the same bit rate, our encoder, being comparable to the H263+ scheme, with a slightly lower (Peak Signal Noise Ratio (PSNR) value, produces a more visually pleasing result. This implementation also preserves scalability of wavelet embedded coding technique. Thirdly, the scheme to handle optimal bit allocation among color bands for still imagery has been modified and extended to accommodate the spatial-temporal sensitivity of the HVS model. The bit allocation among color bands based on Kelly's spatio-temporal CSF model is designed to achieve the perceptual optimum for human eyes. A perceptually tuned, wavelet based, rate scalable video encoding/decoding system has been designed and implemented based on this new bit allocation scheme. Finally to present the potential applications of our rate scalable video codec, a prototype system for rate scalable video streaming over the Internet has been designed and implemented to deal with the bandwidth unpredictability of the Internet.

This study describes the use of Landsat MSS and TM for geological applications in two Scottish areas: Lochindorb and Loch Tuel; and one Malaysian area: Kedah-Perak. The areas are poorly exposed and highly vegetated. The data were digitally processed with the objective of producing more interpretable images. The processes include contrast enhancement, ratioing, subtraction, principal component analysis, discriminant analysis, filtering, the combination of images as colour composites, and producing negative images of the data. Geological interpretation of the most informative images was undertaken by visual interpretation. In the Lochindorb area, Landsat MSS imagery did not prove useful for superficial deposits mapping, and the resolution offers by the TM is still not sufficient for semi-detailed mapping at scale 1: 50,000. The combination of TM imagery and aerial photographs, however, made the mapping task easier and produced "better" map. In the Kedah-Perak area, textural information is more important than spectral information for lithological interpretation and many image units correlate well with major mapped rocks. Lineaments are well expressed on Landsat imagery and are mapped for the Loch Tummel and Kedäh-Perak areas. The lineament maps for both areas confirm many mapped faults and reveal a new prominent lineaments (probably faults). For the Loch Tummel area, the relative merits of TM versus MSS data were examined. Both produced similar results regarding major lineament orientations, but the TM provides a good improvement over the MSS in the ability to map lineaments. For both areas, lineaments appear to be correlated with geomorphology (lithology), and with the occurrence of ore deposits and probably geologic structure for the Kedah-Perak area. Landsat imagery can be used to aid lithological mapping in Malaysia, but has not proved useful for Scotland (U. K. ) because of different objectives and constraints. However, Landsat imagery is an effective tool in mapping lineaments for both areas.

D.Ing. / This thesis deals with an investigation into facial recognition and some variables that influence the performance of such a system. Firstly there is an investigation into the influence of image variability on the overall recognition performance of a system and secondly the performance and subsequent suitability of a neural network based system is tested. Both tests are carried out on two distinctly different databases, one more variable than the other. The results indicate that the greater the amount of variability the more negatively affected is the performance rating of a specific facial recognition system. The results further indicate the success with the implementation of a neural network system over a more conventional statistical system.