A well suited two-dimensional linear recursive filter for image processing

In this paper, a fixed order two-dimensional linear recursive filter is obtained by matching the unit sample response of a fixed-order two-dimensional linear recursive operator with the unit sample response of an ideal circularly symmetric low pass filter. Such an operator thereby obtained possesses the desired circularly symmetric low pass filtering characteristic and is also stable. However, a direct application of this filter to the image processing is generally not well-suited. For this reason, a more appropriate filter called the flipped-filter, which utilizes the circularly symmetric filter, is designed. In order to demonstrate the effectiveness of the flipped-filter in image processing, a set of noise contaminated images are processed by this filter and the processed images are displayed.