ABSTRACT

In this paper a novel medical image processing system is discussed. A modified algorithm for the reduced hardware has been designed. Finally some suitable modification in the hardware is made. In Medical Image Processing System, characterization and recognization of suspended minute particles in cellular fluid is an important problem. The two key steps involved are up-sampling of the region of interest (RoI) and edge detection. In our case, the target RoI’s are already blurred due to the presence of surrounding cellular fluids. So up sampling of these RoIs makes them more blurred and as a consequence, the subsequent edge detection becomes a futile exercise. Henceforth arises the requirement of an interpolation algorithm, which preserves some high pass contents. The B-Spline interpolation satisfies the above-mentioned constraint. In past, few papers dealt with the implementation of B-Spline hardware. However they failed to exploit some important properties of basis functions like periodicity. Here we have efficiently utilized those properties to achieve a novel parallel architecture.