To extract video semantic concepts combing different modalities, a two level multimodal fusion method for video semantic concept analysis is proposed.
为了全面准确地获取视频高层语义信息，提出了一种基于仿生的视频语义分析两级多模式融合算法。
In content-based image retrieval systems, the inconsistency between image low-level features and the concept of high-level expressed by images lead to system semantic gap problem.
在基于内容的图像检索系统中，图像低层特征和图像所表达高层概念之间的不一致性导致系统出现语义鸿沟问题。
The thesis treats coherence as a multi-level concept. And it is a comprehensive occurrence which can be shown in semantic level, pragmatic level, psychological level, cognitive level and etc.
本论文认为连贯是一个多层次性的概念，它是可以体现于语义层面、语用层面、心理层面、认知层面等各个层面的综合。
应用推荐