欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
为了提高图像检索精度，文章在谱哈希的基础上引入最小量化误差的思想，提出了一种基于改进谱哈希的大规模图像检索算法，该算法避免了谱哈希中要求的数据服从均匀分布的假设，并且能够保持数据在原始空间的相似性；引入Boosting算法来确定阈值，使得该算法具有更强的适应性和更广泛的应用；在公开的图像数据集上做了实验，实验结果表明该方法优于谱哈希、局部敏感哈希和迭代量化等哈希算法。
英文摘要：
In this paper, a large-scale image retrieval method based on improved spectral hashing is proposed by introducing the quantization error minimization to improve the image retrieval accuracy. The algorithm gets rid of the assumption that data is uniformly distributed required by the spectral hashing, and can keep the data similarity in the original space. Boosting algorithm is used to determine the threshold so that the algorithm can be more adaptive and more widely used. The algorithm is evaluated on the public datasets, and the experimental results show that the proposed method is better than spectral hashing（SH）, locality-sensitive hashing （LSH） and iterative quantization hashing algorithm.
同期刊论文项目
同项目期刊论文
期刊信息
位置：