We present a novel perspective on characterizing the spectral correspondence between nodes of the weighted graph with application to image registration.
本文提出了一种新的基于谱图矩阵扰动分析的权图顶点对应的图像配准方法。
A novel algorithm for point correspondence is proposed, which combines graph spectral analysis and partial characteristics of the point together via the method of probabilistic relaxation.
文章提出了一种将谱图理论、特征点的局部特征和概率松弛法相结合的特征点匹配算法。
应用推荐