基于局部敏感哈希算法和神经网络学习的跨媒体检索方法

为了提高跨媒体检索的效率,可行的方法是降低数据集中不相关内容的比例。采用局部敏感哈希算法将图像数据映射到汉明空间并利用神经网络学习将文本数据映射到汉明空间的哈希函数,提出一种可以显著提高数据集中相关文件比例的高效跨媒体检索方法。实验结果表明,提出的方法能够有效去除数据集中的不相关内容,相对于已有的跨媒体检索方法,其有效提高了检索效率与准确率。

Fast Cross-Media Retrieval with Locality-Sensitive Hashing and Neural Networks

Existed approaches for cross-media retrieval are computationally expensive due to the curse of dimensionality. Especially, these approaches query in the whole database which contains massive documents unrelated to the query. To efficiently retrieve in multimodal data, it is essential to reduce the proportion of irrelevant documents. A novel cross-media retrieval approach is proposed based on locality-sensitive hashing and neural networks to reduce the proportion of irrelevant documents. The experiment has shown that in the set of the queries'' near neighbors obtained by the proposed method, the proportion of relevant documents can be much improved, indicating that the retrieval based on near neighbors can be effectively performed. Assessments on two public datasets also demonstrate the efficacy of the proposed retrieval method when compared to the baselines.