| 聂秀山,刘兴波,袭肖明,尹义龙.基于相似度驱动的线性哈希模型参数再优化方法.软件学报,2020,31(4):0 |
| 基于相似度驱动的线性哈希模型参数再优化方法 |
| Model Parameter Re-optimization for Linear Hashing Based on Similarity Drive |
| 投稿时间：2019-03-09 修订日期：2019-07-11 |
| DOI：10.13328/j.cnki.jos.005918 |
| 中文关键词: 内容检索 哈希学习 线性模型 参数优化 相似度驱动 |
| 英文关键词:content retrieval hash learning linear model parameter optimization similarity drive |
| 基金项目:国家自然科学基金（61671274，61876098） |
|
| 摘要点击次数: 140 |
| 全文下载次数: 0 |
| 中文摘要: |
| 哈希学习通过设计和优化目标函数，并结合数据分布，学习得到样本的哈希码表示.在现有哈希学习模型中，线性模型因其高效、便捷的特性获得广泛应用.针对线性模型在哈希学习中的参数优化问题，提出一种基于相似度驱动的线性哈希模型参数再优化方法.该方法可以在不改变现有模型各组成部分的前提下，实现模型参数的再优化，提升模型检索性能.该方法首先通过运行现有哈希算法多次，获得训练集的多个哈希码矩阵，然后基于相似度保持度量标准和融合准则对多个哈希码矩阵进行优化选择，获得训练集的优化哈希矩阵，最后利用该优化哈希矩阵对原模型的参数进行再优化，进而获得更优的哈希学习算法.实验表明，该方法对不同的哈希学习算法性能都有较为显著的提升. |
| 英文摘要: |
| Hash learning is first to design an objective function, and then learn the hash codes based on the distribution of samples. In the existing hashing models, linear model is widely used due to its conciseness and high efficiency. For the parameter optimization of linear hashing model, we propose a model parameter re-optimization method based on similarity drive, which can improve the precision of the existing linear model-based hashing algorithms. Given a hashing method, we first run this method for several times with obtaining several hash matrices. Then, we select some bits for these hash matrices to obtain a new final hash matrix based on the similarity preserving degree and a fusion strategy. Finally, we use this new hash matrix to re-optimize the model parameters, and a better hash model is obtained for out-of-sample extension. Extensive experiments are performed based on three benchmark datasets and the results demonstrate the superior performance of the proposed framework. |
|
HTML 下载PDF全文 查看/发表评论 下载PDF阅读器 |
