VLSI Implementation of Barrel Distortion Correction in Endoscopic Images based on Least Squares Estimation

Abstract

An efficient VLSI Implementation of Barrel Distortion Correction (BDC) in Endoscopic Images based on Least Squares Estimation is presented in this paper. Computational complexity is reduced by employing Odd order polynomial, as an approximation to Back-mapping expansion polynomial. This polynomial can be solved in monomial form, by Estrin's algorithm. In Estrin’s algorithm, a high order expression can be factorized in to sub-expression, which can be evaluated in parallel. In our simulation, on comparison with some existing distortion correction techniques, 75% of hardware cost and 70% of memory requirement is reduced by using TSMC 0.18µm technology.