基于分治思想粗匹配和精微匹配相结合的跨模态检索算法

Based on divide and conquer thought combining coarse matching and fine matching cross-modal retrieval algorithm

跨模态检索是人工智能领域的一个重要研究方向,在社会生活中应用广泛,有着巨大的应用价值和经济价值。随着深度学习的兴起,跨模态检索也取得了长足发展。本文借鉴了分治思想和混合推荐的方法,在一个算法框架中构建两个检索模型,分别负责粗匹配和精微匹配。通过特征值取平均值的方式将两个检索模型整合在一起,通过同时使用两个检索模型的检索能力来提升算法的检索效果,增强算法的抗干扰性。

Cross-modal retrieval is an important research direction in the field of artificial intelligence.It is widely used in social life and has huge application value and economic value.With the rise of deep learning,cross-modal retrieval has also made great progress.This paper draws on the idea of divide and conquer and the method of mixed recommendation.Two retrieval models are constructed in an algorithm framework,which are responsible for rough matching and fine matching respectively.The two retrieval models are integrated by averaging feature values,and the retrieval capabilities of the two retrieval models are used at the same time to improve the retrieval effect of the algorithm and enhance the anti-interference ability of the algorithm.