摘要
[目的/意义]数字人文领域中图像包含大量的重要语义信息,是重要的研究对象,然而由于"语义鸿沟"的问题,目前常用的图像自动化标注方法并不适用于这一领域的图像语义描述工作。[方法/过程]提出一种面向数字人文的图像语义描述模型。根据用户认知特征制定图像语义结构化描述框架,并利用多方法组合进行语义特征词抽取与映射。最后,采集中国国家图书馆老照片数据进行实验。[结果/结论]实验结果显示,文章构建的模型在充分符合反映用户认知习惯的同时,对于语义特征词的识别与映射效果较好,有助于解决数字人文环境下图像语义描述问题。
[ Purpose/significance ] In the study of digital humanities, as image contains important semantic information, it becomes an important research subject. However, because of the ＂semantic gap＂ problem, the current commonly used image se- mantic annotation automation method is not applicable to image description in the field of digital humanities. [ Method/process] A semantic description model for digital humanities is proposed. According to the cognitive characteristics of users, the semantic struc- ture description framework is established, and the semantic feature words are extracted and mapped by the combination of multiple methods. Finally, an experiment is carried out by collecting the data of an old photo from the National Library of China. [ Result/ conclusion ] The results show that this model can fully reflect users＇ cognitive habits, at the same time, identifies and maps the se- mantic features of words with better effect, which can help to solve the image semantic description problem in digital humanities en- vironment.
出处
《情报理论与实践》
CSSCI
北大核心
2018年第1期116-121,共6页
Information Studies:Theory & Application
基金
国家自然科学基金项目"云环境下智慧图书馆移动视觉搜索模型与实现研究"的成果之一
项目编号:71673203
关键词
数字人文
图像
语义描述
模型设计
digital humanities
image
semantic description
model design
