视频重着色可以有效地提高色盲用户对多媒体信息的可访问性,但是目前视频重着色研究存在颜色不一致性问题.针对该问题,提出一种基于层次化结构信息的视频重着色的方法.该方法在视频层次化结构中应用基于CIELAB颜色空间旋转的重着色方法,采用颜色空间量化和哈希表相结合的方式实现实时性,同时既使每个镜头中的颜色变换函数保持相同,又使场景中镜头之间的颜色变换函数变化缓慢,实现场景内颜色的一致性.实验结果表明,对于不同种类的视频,该方法都能有效地使视频更好地被色盲用户感知,为视频重着色在多媒体搜索中的应用起到推动作用.
Video re-coloring is an effective way to help the color-blind access visual content. However, previous work does not take the color inconsistency into account. In this paper, a novel video re-coloring method has been proposed to solve the color inconsistency problem. In our solution, the shot and the scene information are integrated into a hierarchical structure. To map color spaces in real- time,the CIELAB color space conversion, RGB color space quantization and the hashing method have been employed. These tech- niques lead to an identical conversation function in a shot, as well as gradual changing function in the same scene. Experiment is con- ducted on different kinds of videos, and the results suggest that the proposed method not only helps colorblind users videos perceive the video better, but also shows the possibility of promoting the re-coloring application in multimedia search.