Color Calibration Model in Imaging Device Control using Support Vector Regression

Abstract

In the color system of a computer, the nonlinearity of the image acquisition device and the display device may result in the difference between the colors displayed on the screen and the actual color of objects, which requires for color correction. This paper introduced the Support Vector Regression (SVR) to establish a color correction model for the nonlinear imaging system. In the modeling process, the Successive 3σ Filter was used to eliminate the large errors found in the color measurement. Because the SVR model of RBF kernel has two important parameters (C, γ) that need to be determined, this paper applied Least Mean Squared Test Errors Algorithm to optimize the parameters to get the best SVR model. Compared with quadratic polynomial regression, BP neural network and relevance vector machine, SVR has better performance in color correction and generalization.

DOI: http://dx.doi.org/10.11591/telkomnika.v11i10.3336