Ant Improvement of Just Noticeable Color Difference Estimation

In this paper, the estimation of just noticeable color difference in color images is improved by using a new spatial masking The internal generative mechanism of human of brain theory implies that the human visual system (HVS) is sensitive to the orderly stimulus possessing structural regularity which is easily to be predicted and is insensitive to the disorderly stimulus containing structural irregularity. It is obviously that the spatial masking in color images may be overestimated in the region with orderly structures and underestimated in the region with disorderly structures. By using a simple prediction model imitating the brain works of the HVS, the structural irregularity is computed to build a new masking function that can be used to improve the estimation of just noticeable color difference for color images. The masking function is further extended to build a color visual model of estimating the visibility thresholds of color images for performance comparison. Simulation results demonstrate that the proposed method is able to obtain better performance of estimating just noticeable color difference.