In this paper, a novel motion texture approach is presented for synthesizing long character motion (e.g., kungfu) that is similar to the original short input motion. First, a new motion with repeated frames is generat...In this paper, a novel motion texture approach is presented for synthesizing long character motion (e.g., kungfu) that is similar to the original short input motion. First, a new motion with repeated frames is generated by exploiting the symmetric properties of the frames and reversing the motion sequence playback in a given motion sequence. Then, the order of the above motion sequence is rearranged by putting the start and the end frames together. The graphcut algorithm is used to seamlessly synthesize the transition between the start and the end frames, which is noted as graphcut-based motion-texton. Finally, we utilize the motion-textons to synthesize long motion texture, which can be patched together like the image texture synthesis method using graphcut algorithm, and automatically form a long motion texture endlessly. Our approach is demonstrated by synthesizing the long kungfu motion texture without visual artifacts, together with post-processing including our new developed graphcut-based motion blending and Poisson-based motion smoothing techniques.展开更多
提出了一种基于梯度约束SLIC(simple linear iterative clustering)的快速视频目标提取方法,允许在关键视频帧上提供少量用户交互下,该方法能够快速并精确地提取复杂视频片段中的视频目标。首先,采用梯度约束的SLIC算法对视频片段进行...提出了一种基于梯度约束SLIC(simple linear iterative clustering)的快速视频目标提取方法,允许在关键视频帧上提供少量用户交互下,该方法能够快速并精确地提取复杂视频片段中的视频目标。首先,采用梯度约束的SLIC算法对视频片段进行预处理,有效降低待处理的视频数据量;其次,以预处理生成的超像素为结点构建三维无向图,在此基础上定义能量函数,并结合外观特征与运动特征建立鲁棒的相似外观度量机制;最后,采用最大流/最小割算法最小化能量函数以得到三维无向图的最优划分,从而最终实现视频目标提取。实验结果表明,该方法在处理包含复杂场景的视频片段时能够获得理想的视频目标提取结果,且时间效率相比现有视频目标提取方法明显提高。展开更多
基金Project supported by the National Natural Science Foundation of China (Nos. 60573153 and 60533080), and Program for New Century Excellent Talents in University (No. NCET-05-0519), China
文摘In this paper, a novel motion texture approach is presented for synthesizing long character motion (e.g., kungfu) that is similar to the original short input motion. First, a new motion with repeated frames is generated by exploiting the symmetric properties of the frames and reversing the motion sequence playback in a given motion sequence. Then, the order of the above motion sequence is rearranged by putting the start and the end frames together. The graphcut algorithm is used to seamlessly synthesize the transition between the start and the end frames, which is noted as graphcut-based motion-texton. Finally, we utilize the motion-textons to synthesize long motion texture, which can be patched together like the image texture synthesis method using graphcut algorithm, and automatically form a long motion texture endlessly. Our approach is demonstrated by synthesizing the long kungfu motion texture without visual artifacts, together with post-processing including our new developed graphcut-based motion blending and Poisson-based motion smoothing techniques.
文摘提出了一种基于梯度约束SLIC(simple linear iterative clustering)的快速视频目标提取方法,允许在关键视频帧上提供少量用户交互下,该方法能够快速并精确地提取复杂视频片段中的视频目标。首先,采用梯度约束的SLIC算法对视频片段进行预处理,有效降低待处理的视频数据量;其次,以预处理生成的超像素为结点构建三维无向图,在此基础上定义能量函数,并结合外观特征与运动特征建立鲁棒的相似外观度量机制;最后,采用最大流/最小割算法最小化能量函数以得到三维无向图的最优划分,从而最终实现视频目标提取。实验结果表明,该方法在处理包含复杂场景的视频片段时能够获得理想的视频目标提取结果,且时间效率相比现有视频目标提取方法明显提高。
