CAJ | 학술논문

图结构特征的提取以及距离度量是计算机视觉和模式识别中的重要内容。针对传统的图上下文描述子中只考虑每个柱形区域内的特征点数目而忽略特征点之间的结构信息的问题，提出一种图的结构上下文描述子及距离度量方法。首先对图的所有顶点建立图结构上下文描述子；其次，利用二次卡方（QC）距离方法实现图的距离度量；最后，采用最小生成树聚类算法实现图聚类。实验表明，该方法对于一些非刚性变换的图聚类有较好的效果。
도결구특정적제취이급거리도량시계산궤시각화모식식별중적중요내용。침대전통적도상하문묘술자중지고필매개주형구역내적특정점수목이홀략특정점지간적결구신식적문제，제출일충도적결구상하문묘술자급거리도량방법。수선대도적소유정점건립도결구상하문묘술자；기차，이용이차잡방（QC）거리방법실현도적거리도량；최후，채용최소생성수취류산법실현도취류。실험표명，해방법대우일사비강성변환적도취류유교호적효과。
Feature extraction and similarity measurement for graphs are important issues in computer vision and pattern recognition .However,traditional graph context method only considered the number of feature points and ignored the structure information of feature points in each quadrant area,so a new graph struc?ture context(GSC)descriptor and its similarity measurement method were proposed .Firstly,graph structure context descriptor was presented based on the feature point set .Secondly,graph distance was measured by using Quadratic Chi-square (QC) distance. Finally,minimum spanning tree clustering algorithm was adopt?ed to realize graph clustering .The experimental results demonstrate that the proposed method performs better for the graphs under some non-rigid transformation.