A Novel Algorithm Based on Improved BP Neural Network and Its Application in Color Management

Color management plays a key roles in the color image reproduction in computer multimedia processing. The paper advances a novel color management algorithm through analyzing printer color rendering principle in order to realize the accurate color management for color printer. First, standard color target for scanner color correction is taken for experimental sample and color blocks in color shade district is substituted for complete color space to solve the difficulties of experimental color blocks selecting. Second, combining the advantages of Legendre wavelet and BP neural network, the paper designs a new BP neural network model in which Legendre wavelet is used to speed up the convergence of BP model and use the model to converse color space for color printer, Finally the experimental results show that the model can improve color conversion accuracy and shorten calculation time effectively compared with ordinary BP network model.