Content-color-dependent screening (CCDS) using regular or irregular clustered-dot halftones

In our previous work, we have presented an HVS-based model for the superposition of two clustered-dot color halftones, which are widely used for electrophotographic printers due to their relatively poor print stability. The model helps us to decide what are the best color assignments for the two regular or irregular halftones that will minimize the perceived error [1]. After applying our model to the superposition of three and four clustered-dot color halftones, it was concluded that this color assignment plays a significant role in image quality. Moreover, for different combinations of colorant absorptance values, their corresponding best color assignments turn out to be different. Hence, in this paper we propose to apply different color assignments within the image depending on the local color and content of the image. If the image content locally has a high variance of color and texture, the artifacts due to halftoning will not be as visible as the artifacts in smooth areas of the image. Therefore, the focus of this paper is to detect smooth areas of the image and apply the best color assigments in those areas. In order to detect smooth areas of the image, it was decided to segment the image based on the color of the content. We used the well-known K-means clustering algorithm along with an edge detection algorithm in order to segment an image into clusters. We then used our spatiochromatic HVS-based model for the superposition of four halftones in order to search for the best color assignment in a particular cluster. This approach is primarily directed towards good quality rendering of large smooth areas, especially areas containing important memory colors, such as flesh tones. We believe that content-color-dependent screening can play an important role for developing high quality printed color images.