Improving Color Text Sharpness in Images with Reduced Chromatic Bandwidth

Current digital imaging systems often reduce the chroma bandwidth to speed transmission or save storage. While has become routine for most video applications, it is a becoming true of digital camera systems. Chroma bandwidth is usually reduced by converting the RGB ima into a single luminance and two color difference image and then subsampling the color difference images. T motivation lies in the reduced bandwidths of the visu system for the color signals, as evidenced by contr sensitivity function (CSF) data. Systems using chroma subsampling can exhibit visual equivalence to otherw identical non-subsampled systems, despite the phys color blur introduced by subsampling. While thi equivalence holds true for images of real scenes a optically captured scenes, problems arise with synthetica generated imagery, such as color geometrics and type fo The possible sources of these problems include: 1) devia from natural image chromatic power spectra, 2) nonline masking effects between the luminance and chromina visual mechanisms, and 3) filtering processes in t luminance-chrominance domain create signals outside the RGB gamut. By taking advantage of nonline perceptual effects and the gamut clipping issues, we h developed a simple point-processing algorithm th accomplishes a reduction in the visibility of chromatic blur