1) middle-level feature
中级特征
1.
It gets the candidate boundary points through shot detection,extracts multi-modal middle-level features at the boundary points such as visual,audio type,motion and caption to generate a feature set as the input of the model,and uses trained Nave Bayes model to classify the candidate boundary points.
通过对新闻视频进行镜头检测,获得候选故事边界点,从候选边界点周围镜头提取多模态中级特征,形成属性集合作为输入,应用朴素贝叶斯模型对候选边界点进行分类后对结果进行后处理,得到新闻故事。
3) quantative feature
量级特征
1.
Moreover,the author points out that the quantative features of Γ-distribution lie in that the less the larger, the more the less, and extraordinary vast is less.
文章通过对新疆断裂构造的概申分布拟合优度检验，证实了断裂在长度上与间距上服从Γ-分布，并进一步讨论了肝Γ-布的量级特征及其与灾变事件的关系。
5) low-level feature
低级特征
1.
This paper introduces the concepts of a content-based image retrieval technology and medical image retrieval system,analyzes the present research situation of medical image retrieval system,and by using multiple aspects normalized image retrieval method and the combination of low-level features and semantic features,establishes a medical image retrieval system based on ontology.
介绍了基于内容的图像检索技术和医学图像检索系统的相关概念,分析了医学图像检索系统的研究现状,采用多特征归一化图像检索方法、低级特征和语义特征相结合的二级检索机制,构建了一个基于本体的医学图像检索系统。
2.
Although a variety of retrieval techniques have been designed to reduce the number of the candidate images,most techniques are based on low-level features and consider little or no high level semantic information.
虽然许多不同的检索技术被设计用于减少需要检索的目标图像的数量,可是绝大多数检索技术都是基于低级特征并且没有或很少考虑高级语义信息。
6) utterance level feature
段级特征
1.
This method chooses the fittest utterance level feature length according to the experiments.
提出了一种使用段级语音特征对测试进行说话人分段从而实现对话环境下说话人分段算法,算法实现中基于车比雪夫和不等式提出了基于协方差模型的段级特征的距离测度描述。
补充资料：中级
1.中层。 2.介于高级和低级之间的级别。
说明：补充资料仅用于学习参考，请勿用于其它任何用途。
参考词条