该算法首先通过对边缘密度图进行投影分析进行自顶向下的粗定位，然后在此基础上利用垂直边缘的连接强度进行自底向上的精确定位。
The algorithm first makes a top-down coarse location using the projection analysis of the edge density graph and then makes a bottom-up precise location based on the vertical edge linkage intensity.
为了提高二值图像的压缩比，采用自顶向下的编码方法，提出一种基于等腰直角三角形的分形图像压缩算法，并进行仿真实验。
In order to improve the ratio of binary image compression, using top-dowm method of coding, bring forward a fractal image compression method based on isosceles right triangle, and simulated.
本文以归纳程序综合系统NDIPS为背景，介绍自顶向下的归纳程序综合算法，以及程序假设空间的组织、搜索所采用的技术和策略。
With the inductive program synthesis system NDIPS, this paper presents its synthesis algorithm, the organization of its hypothesis space and the search strategies it adopted.
目前存在的决策树归纳算法大多数是基于自顶向下的贪婪算法，它在每个结点都执行一个局部最优决策。
The existing majority of the decision tree inductions are based on a top-down greedy algorithm, which make a locally optimal decision at each node.
实验证明，与现有的自顶向下挖掘算法相比，该算法是快速而有效的。
Comparing with presented top-down mining algorithms, the experiments indicate that it is fast and efficient.
然后根据标签间的相关度度量选出图片结果集中与查询标签高相关的标签集，接着采用一种自顶向下启发式的图划分算法来自动对次相关标签集进行分类。
The candidate tags can be shortlisted according to their tag relevances to the query tags. The shortlisted tags are then clustered on-the-fly using a graph partitioning algorithm.
然后根据标签间的相关度度量选出图片结果集中与查询标签高相关的标签集，接着采用一种自顶向下启发式的图划分算法来自动对次相关标签集进行分类。
The candidate tags can be shortlisted according to their tag relevances to the query tags. The shortlisted tags are then clustered on-the-fly using a graph partitioning algorithm.
应用推荐