摘要
针对基于内容的检索过程中特征提取完成后,多种不同类型的特征如何整合的问题,提出了一种基于属性论方法的偏好学习方法.这种偏好学习方法旨在成为特征整合的依据,为分类、检索、识别等后续的人工智能操作以支撑.以基于内容的图像检索为例,将图像的颜色、纹理、形状和草图等特征通过心理权重整合后,提出了属性坐标分类法,实现了效果较好的局部偏好检索,并证明了所提出方法的有效性.
In solving pattern recognition, pattern classification process, how to integrate a variety of characteristic values was a difficult problem to solve. A new solution was proposed. The "analysis approach of attribute coordinate" in "attribute theory" to record user's behavior in the process of retrieval was applied and was distilled the user's psychological preference, which could provide more satisfactory retrival results for the users. One feedback machanism was built for psychological preference learning, which made the retrieval results steadily approach user's expectation.
出处
《郑州大学学报(理学版)》
CAS
北大核心
2011年第2期66-72,共7页
Journal of Zhengzhou University:Natural Science Edition
关键词
基于内容的图像检索
属性坐标分析法
特征提取
心理偏好
特征整合
content-based image retrieval
traction
psychological preference
feature analysis approach of attribute coordinate
feature ex integration