个性化图像检索和推荐

为了解决信息过载的问题,个性化图像检索和推荐技术成为目前图像检索领域的新趋势,其不仅可提高检索的效率和准确率,还可满足用户的个性化需求.根据不同个性化信息的数据源,可将个性化图像检索和推荐分为基于内容的个性化图像检索和推荐与协同过滤个性化图像检索和推荐.对于基于内容的个性化图像检索和推荐,分析了用户兴趣获取、用户兴趣表示和个性化实现3个核心环节,并对所采用的关键技术进行了对比,指出了优缺点;对于协同过滤个性化图像检索和推荐,分析了基于用户、物品和模型的3种协同过滤方法.最后分析对比了基于内容和协同过滤2种个性化图像检索和推荐方法,并指出了未来的工作方向.

Personalized Image Retrieval and Recommendation

In order to solve the problem of information overload,personalized image retrieval and recommendation technology has become the new trend in image retrieval area.It can not only improve the efficiency and accuracy of retrieval,but also meet users' personalized requirements.Personalized image retrieval and recommendation can be divided into content-based personalization and collaborative-filtering-based personalization according to different sources of personalized information.Firstly the development of image retrieval are summarized.Then the key technologies of content-based personalized image retrieval and recommendation are analyzed from three aspects,user interest acquisition,user interest representation and personalization implementation.The key technologies are compared.Their advantages and disadvantages are pointed out.For collaborative-filtering-based personalized image retrieval and recommendation,user-based,item-based and model-based collaborative filtering methods are contrasted.At the end of the paper,the content-based methods and the collaborative filtering methods are discussed,and the future work is shaped.