A Tabulation-Based Fast Algorithm for 2-Dimensional Convolution

摘要

Building image databases typically requires processing and analysis of huge collections images as a first step. Image filtering is a fundamental operation in image processing. Filtering essentially involves convolution operation. Any scheme that speeds up the convolution which in turn speeds up image filtering operation would be beneficial to the database building process. This paper presents a fast algorithm to perform the 2-dimensional filtering of image data using tables of intermediate values. The simplicity and regularity of the scheme can also be used for hardware implementation.