关键词: 图像语义标注, 标注效果层次性, 图像语义描述框架, 用户信息需求, 敦煌壁画, 数字图像

Abstract:

Users' requirements for information to understand Dunhuang digital frescoes have been obtained through interviews, based on which questionnaires have been made to know how well users understand the frescoes. Finally, comparisons  have been  made between users' understanding level before and after using tagging system. The effectiveness is assessed of hierarchic SDFDI in proper order. It has been found that when perceiving frescoes images, users need  the information of Preiconographic level most of all, and the model performs best in annotating problems of this level, while the model help improve the problems of signal level least, as well as that SDFDI performs better in annotating story frescoes than in those  about buddhist scenes.

Keywords: Image semantic annotation, Semantic annotation effect, Semantic description framework of digital image(SDFDI)；User information requirement, Dunhuangfrescos, Digital image