图像购物搜索技术研究综述

摘要

Traditional electronic commerce websites mainly use the key words of commodity to search for goods. However, text information can’t describe all the characteristics of the commodity, and can cause subjectivity of human operation. The image retrieval technology based on content provides a new opportunity for the development of electronic commerce website by extracting the visual features of the images. This thesis combs the researches on the shopping technology based on image retrieval, summarizes research status, analyses the principle, and puts forward future trend.%传统的电子商务网站主要通过商品名称的关键字来进行商品搜索，然而文字信息难以完整的描述商品的各种特征，并且存在人为操作的主观性，导致搜索结果与用户意图之间存在较大差异。基于内容的图像搜索技术通过提取商品图像的视觉特征进行搜索，为电子商务网站的发展提供了新的契机。本文在此基础上对图像购物搜索技术的研究工作进行梳理，总结研究现状，分析技术原理，并指明未来的发展趋势。