Chroma discrimination for natural images

Image quality is one of the decisive factors for consumers when buying electronic display. Chroma is an important aspect affecting image quality evaluation. The paper investigated the chroma discrimination for natural images. Images were generated by decreasing chroma from original image to different extent via image processing. The JND of chroma was measured by perception experiments. It has been found that the JND of Chroma is about 0.6~0.7 in ΔE₀₀ for most images. In addition, the effect of image content on JND of chroma is not significant. The repeatability of the experiment is also verified. These results can be used to direct the industry in the design of display systems by indicating where to put the limited efforts in order to improve the subjective quality as much as possible.