通过对视频纹理定义的分析,将视频纹理合成转化为一个优化组合问题。提出一种应用分段遗传算法的视频纹理合成算法,采用分段遗传算法,对有限长度的源视频进行加工,得到可无限播放的连续视频序列。算法采用更适当的相似性尺度和测量准则,省去了大量复杂的对源视频的预处理,分段的搜索策略只需要用很少的遗传代数即可快速合成出质量很高的视频纹理。与现有的视频纹理合成方法比较,该算法具有较小的计算复杂度,在合成的速度和质量上都有所提高。另外,实验结果给出了种群大小以及最大遗传代数对合成质量和速度的影响。
Texture synthesis is a very active research area in computer vision and graphics,and video texture synthesis is one subset of it.The goal of this technology is:Giving a video texture sequence,a new video sequence can be synthesized,which is similar to the input one in texture.In process of synthesizing video textures,creating video sequence can be translated into a combinatorial optimum problem.Segment Genetic Algorithm(SGA) is introduced into the process of synthesizing video textures and a new video texture synthesis algorithm is proposed,which is used to generate infinitely continuous streams of video from a finite video clip.High-quality video texture can be synthesized in a few generations due to segment searching strategy of SGA.Compared with representative video texture synthesis algorithms nowadays,this algorithm has less computational complexity and improves the efficiency of synthesis.In addition,how the population size and the max number of generations influence the speed and quality of synthesis is analyzed.