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.
出处
《情报学报》
CSSCI
北大核心
2007年第5期691-698,共8页
Journal of the China Society for Scientific and Technical Information
关键词
风景图片
季节特征
颜色直方图
多元线性回归
natural scenes, season features, color histogram, multiple linear regression
