Successive Blurring Techniques for Image Processing
Anton E. Milev
Institute of Information Technologies, 1113 Sofia
Abstract: A new family of image processing algorithms, called STEC /Standard
Time Equivalent Conversions/ is introduced in this paper . Their main advantage
is their parallel design, which makes them ideal for implementation in parallel
processing systems. The notion of parallelism in STEC is analogous to physical
diffusion processes, which makes them suitable for wide area of tasks beside
image recognition. Further in the paper a new concept called dynamic moments
is examined. It is demonstrated how the latter can be used with contour and
global methods for image recognition. Experimental results and graphics are provided.
Keywords: Image processing algorithms, parallel design, image recognition.