摘要
在图片相似性搜索中,需要将图片特征向量的相似性搜索问题转化为K近邻问题,这就需要了解K近邻问题的定义,以及局部敏感哈希的数学定义。此外,还需引入一个可以用于实际应用的局部敏感哈希的算法,并分析此算法的正确率和算法复杂度。
To solve the image similarity identification problem, the similarity search of feature vector must be converted to K- Near- est Neighbor problem. K- Nearest Neighbor problem is then introduced, and then the mathematical definition of Locality- sensitive Hashing is introduced. An applied algorithm of Locality- sensitive Hashing was described, with the analysis of the rate of accuracy and time complexity.
出处
《上饶师范学院学报》
2013年第6期76-79,共4页
Journal of Shangrao Normal University
关键词
LSH
K近邻
高维向量
搜索
LSH
K- Nearest Neighbor
High Demensions
Search
