节点文献
基于图像序列的非刚体三维运动恢复
NON-RIGID OBJECT 3-D MOTION RESTORATION BASED ON IMAGE SEQUENCES
【摘要】 研究图像序列中非刚体的三维运动重建问题。介绍非刚体运动重建的两种重要算法:奇异值分解法和线性迭代法。对迭代算法所用的重构方法进行修改,在迭代过程中应用结果更为精确的重构方法来求解非刚体的模型和旋转矩阵。对真实图像序列的实验结果验证了该算法的有效性和精确性。
【Abstract】 This paper addresses the issue of restructuring 3 dimensional motion of non-rigid object in image sequences.Two important algorithms of non-rigid object motion restructure are introduced,namely the SVD and the linear iterative method.The reconstructing method,which is used in iterative algorithm,has been improved;the reconstructing method with more accurate results is used to recover the model and motion matrix of non-rigid object in iterative process.Experiment results on real image sequences validate the effectiveness and accuracy of our method.
【关键词】 非刚体;
图像序列;
由运动恢复结构;
奇异值分解;
线性迭代;
【Key words】 Non-rigid object Image sequences Restoring structure from motion Singular value decomposition(SVD) Linear iterative;
【Key words】 Non-rigid object Image sequences Restoring structure from motion Singular value decomposition(SVD) Linear iterative;
【基金】 国家自然科学基金项目(60873020);浙江省自然科学基金项目(Z1080702)
- 【文献出处】 计算机应用与软件 ,Computer Applications and Software , 编辑部邮箱 ,2011年09期
- 【分类号】TP391.41
- 【被引频次】6
- 【下载频次】153