彩色自然风景图片的四季特征提取  

Extraction of Season Features in Natural Scenes for Retrieval

作　　者：黄琨[1] 赖茂生[2] 

机构地区：[1]北京师范大学管理学院,北京100875 [2]北京大学信息管理系,北京100871

出　　处：《情报学报》2007年第5期691-698,共8页Journal of the China Society for Scientific and Technical Information

摘　　要：根据彩色自然风景图片的特点,提出＂天空去除＂和＂1/2区域分析＂改进法提取图像的颜色特征,同时采用四季轮盘调查法收集用户评价,然后通过多元线性回归方法建立颜色特征与用户评价的映射关系,用于彩色自然风景图片四季特征的自动提取.通过实验验证了该映射机制对于正确预测彩色自然风景图片四季特征的有效性.According to the features of colorful natural scenes, it uses sky exclusion plus 1/2 area analysis to extract images＇ color features. At the same time, it carries out an investigation to collect users＇ remarks about images. Then, based on these two kinds of features, it establishes the mapping between color and subjective impressions by multiple linear regression, which could be used for extracting season features automatically. Finally, the mapping is also testified valid for forecasting and indexing the season features correctly.

关 键 词：风景图片 季节特征 颜色直方图 多元线性回归 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]