Intelligent decoupled SAC-SVD method in color space transformation of computer vision

Abstract

This paper proposes a decoupled SAC-SVD method to calculate color space transformation in single assignment C (SAC) language. Chromaticity coordinate transformation is discussed based on the decoupled SAC-SVD method. The conventional color space transformation is decomposed into three independent equations, respectively. Singular value decomposition (SVD) transformation is proposed to provide the constant values for the associated derivation. Such an algorithm is suitable for the arithmetic operation of single assignment C. The hardware description language VHDL can be generated by using Data Flow Graph in a single assignment C compiler. The results will show that the proposed algorithm can shorten the execution time to a certain extent. It is evident that the proposed intelligent algorithm can be very helpful in implementing the hardware-interface or software-driver for the computer vision. © 2009 Springer Berlin Heidelberg.