Interpolating algorithm optimization and FPGA implementation in image scaling engine

Abstract

Bi-cubic interpolation algorithm is commonly used in image scaling, but traditional cubic interpolation has its own shortcomings such as complicated computation, long computational time and so on. For these problems, the paper studies traditional cubic kernel function and proposes an optimized algorithm with adjustable coefficients. This algorithm utilizes an modifying coefficient lambda to amend the coefficients in the kernel function, which helps the scaling system choose best algorithm with different images. Then, the superiority is verified by MATLAB simulation and the optimized algorithm is applied to the image scaling engine called scaler through the verification in FPGA.