The work presents some new algorithms realized recently in the package TESTAS. The package decides whether or not DFA is synchronizing, several procedures find relatively short synchronizing words and a synchronizing ...The work presents some new algorithms realized recently in the package TESTAS. The package decides whether or not DFA is synchronizing, several procedures find relatively short synchronizing words and a synchronizing word of the minimal length. We check whether or not a directed graph has a road coloring that turns the graph into a synchronizing deterministic finite automaton (DFA). The algorithm finds the coloring if it exists. Otherwise, the k-synchronizing road coloring can be found. We use a linear visualization of the graph of an automaton based on its structural properties.展开更多
传统的图像聚类方法存在对初始数据敏感且计算复杂度高的问题,且图像全局特征难以有效地表达图像内容。针对这些问题,提出一种基于Union-Find的图像聚类方法。首先,该方法采用视觉词袋模型Bo VWM(Bag of Visual Words Model)来描述图像...传统的图像聚类方法存在对初始数据敏感且计算复杂度高的问题,且图像全局特征难以有效地表达图像内容。针对这些问题,提出一种基于Union-Find的图像聚类方法。首先,该方法采用视觉词袋模型Bo VWM(Bag of Visual Words Model)来描述图像内容并且利用投票方法来计算每对图像的相似度得分;然后,对于相似度得分大于给定阈值的图像对进行union和find两个操作并将相连的分量形成聚类结果。实验结果表明,该方法较之于传统方法能较好地改善图像聚类效果,且不需要初始聚类数目作为先验参数。展开更多
文摘The work presents some new algorithms realized recently in the package TESTAS. The package decides whether or not DFA is synchronizing, several procedures find relatively short synchronizing words and a synchronizing word of the minimal length. We check whether or not a directed graph has a road coloring that turns the graph into a synchronizing deterministic finite automaton (DFA). The algorithm finds the coloring if it exists. Otherwise, the k-synchronizing road coloring can be found. We use a linear visualization of the graph of an automaton based on its structural properties.
文摘传统的图像聚类方法存在对初始数据敏感且计算复杂度高的问题,且图像全局特征难以有效地表达图像内容。针对这些问题,提出一种基于Union-Find的图像聚类方法。首先,该方法采用视觉词袋模型Bo VWM(Bag of Visual Words Model)来描述图像内容并且利用投票方法来计算每对图像的相似度得分;然后,对于相似度得分大于给定阈值的图像对进行union和find两个操作并将相连的分量形成聚类结果。实验结果表明,该方法较之于传统方法能较好地改善图像聚类效果,且不需要初始聚类数目作为先验参数。