high performance computing on graphics processing units: hgpu.org

Digital image halftoning is a widely used technique. However, achieving high fidelity tone reproduction and structural preservation with low computational time-cost remains a challenging problem. This paper presents a highly parallel algorithm to boost the real-time application of the serial structure-preserving error diffusion. The contrast-aware halftoning approach is one such technique with superior structure preservation, but offers limited opportunity for GPU acceleration. In this paper, our method integrates the contrast-aware halftoning into a new parallelizable error-diffusion halftoning framework. To eliminate visually disturbing artifacts resulting from the parallelization, we propose a novel multiple quantization model and the space-filling curve to maintain the tone consistency, blue noise property and structure consistency. Our GPU implementation on a commodity PC platform achieves a real-time performance for a moderate-sized image. We demonstrate high quality and performance of the proposed approach with a variety of examples, and provide comparisons with the state-of-the-art methods.