No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

Corners provide strong evidence for the location of objects. Therefore, multi-scale corner detection is very important in image processing, as it provides a sound balance between noise removal and preserving detail. The robustness of corner detection has been used in many existing applications in object recognition and interpretation. This thesis aims to analyse and design a multi-scale corner detector. A simple gradient expression in scale-space that describes the V-, T- and X-type corners in a universal model is defined. We also describe a corner detector, based on Moments of the Gradient in Scale-space (MoGS). The response of this detector is proportional to the edge intensity difference and the sine of the aperture of the corner. The localisation of this algorithm for V- and X-type corners is invariant to scales. The computation of corner attributes for edge orientation, aperture, contrast, corner type and size are evaluated. This algorithm and other corner detectors are evaluated using synthetic and natural images. The results show that the MoGS operator is superior to the Plessey, Kitchen and SUSAN (extended to consider scale) operators in corner detection and localization in scale-space.

Current automatic speech recognition systems make the assumption that all the articulators in the vocal tract move in synchrony with one another to produce speech. This thesis describes the development of a more realistic model that allows some asynchrony between the articulators with the aim of improving speech recognition accuracy. Experiments on the TEVHT database demonstrate that higher phone recognition accuracy is obtained by separate modelling of the voiced and voiceless components of speech by splitting the speech spectrum into high and low frequency bands. To model further articulator asynchrony in speech production requires a representation of speech that is closer to the actual production process. Formant frequency parameters are integrated into typical Mel-frequency cepstral coefficient representation and their effect on recognition accuracy observed. The formant frequency estimates can only accurately be made when the formants are visible in the spectrum, so a technique is developed to ignore frequency estimates generated when the formants are not visible. The formant data allows a unique method of vocal tract normalization, which improves recognition accuracy. Finally a classification experiment examines the potential improvement in speech recognition accuracy of modelling asynchrony between the articulators by allowing asynchrony between all the formants.