Study On Contrast Sensitivity Function Of Human Color Vision Under Uniform-scaled Color Space-Principle, Testing And Application

Frequency response is one of the important characteristics of human vision, which worksas a perceptual information capture system. This characteristic is usually described as ContrastSensitivity Function (CSF). It is significant for processing, compression, quality evaluation andcommunication of color image. The existing research results are based on neutral gray color andon the opposite color space YCbCr(Luminance/Blue-Yellow/Red-Green).Based on uniform-scaled color space CIE1976L*a*b*(CIELAB) and supported by threevariables of Lightness (L)/Chroma(C)/Hue (H), visual CSF in LCH color space are tested.According to the visual tested data, a mathematical model for the CSF of L/C/H is constructedand applied in color image processing, such as image filtering, reduction of spectral data, etc.The results of research and main innovation are as follows:1. Based on analysis of color spaces, under medium levels of lightness and chroma, CSF ofLightness (CSFL), CSF of Chroma (CSFCH) and CSF of Hue (CSFH) are tested and obtained inLCH color space for the first time. The corresponding mathematical model is constructed.2. Under the basic conditions of medium levels of lightness (L=50) and chroma (CH=50),24groups of CSF data are obtained from the visual tests, i.e.: six CSFLunder six different hueangles, six CSFCHunder six different hue angles and twelve CSFHunder twelve different hueangles.3. Close associated with the tolerances of human visual color discrimination, underabove-mentioned medium levels of lightness and chroma, the differences among CSFLundersix different hue angles are not obvious, while the differences among CSFCHunder six differenthue angles, and the differences among CSFHunder twelve different hue angles are evident. Thevalues of CSFLare higher than the values of CSFCHand CSFH, while the CSF with highervalues possesses also wider spectral range.4. A frame processing flow is built up for visual filtering of color image. The starting pointof the flow is an image under CIELAB color space. After a transformation from CIELAB intoLCH color space, the image is transformed into spectral domain, and then it is filtered usingCSFL/CSFCH/CSFH. After the visual filtering, an inverse transform is implemented and the LCHimage is obtained. The LCH image can be transformed into CIELAB color space. The flowbuilds a connection of visual frequency response with uniform-scaled color space CIELAB. In conclusion, the research of CSF, which connected with uniform-scaled color space, ishelpful to expand the knowledge of human visual characteristics, its application in the area ofimage science and engineering can improve the color image processing based on colorimetricspace.