Colorblind-Shareable Videos by Synthesizing Temporal-Coherent Polynomial Coefficients

摘要

To share the same visual content between color vision deficiencies (CVD)and normal-vision people, attempts have been made to allocate the twovisual experiences of a binocular display (wearing and not wearing glasses)to CVD and normal-vision audiences. However, existing approaches onlywork for still images. Although state-of-the-art temporal filtering techniquescan be applied to smooth the per-frame generated content, they may fail tomaintain the multiple binocular constraints needed in our applications, andeven worse, sometimes introduce color inconsistency (same color regionsmap to different colors). In this paper, we propose to train a neural networkto predict the temporal coherent polynomial coefficients in the domain of global color decomposition. This indirect formulation solves the colorinconsistency problem. Our key challenge is to design a neural network topredict the temporal coherent coefficients, while maintaining all requiredbinocular constraints. Our method is evaluated on various videos and allmetrics confirm that it outperforms all existing solutions.