Mask decompositions for digital image processing

The purpose of this study was to investigate the properties of mask convolvers as image processing tools, and to look at sensible approximations to facilitate real time applications. The characteristics of real time image processing systems were briefly reviewed. Afterwards, a theoretical study was made of spatial operators, and new theorems were developed to establish some important properties of mask convolvers. These results were used to obtain a new algorithm to reduce the effective size of mask operators. Some results that show the performance level of the algorithms developed are presented. The quality of these results, as well as the simplicity of the algorithms themselves, suggest that the approach taken is a promising one and that further work in this direction would be fruitful.