摘要
针对图像检索中基于单一全局颜色特征或局部特征的检索方法存在查准率和查全率低等问题,提出了一种融合信息熵和改进尺度不变特征变换算法的图像检索方法。首先,利用改进的尺度不变特征变换算法提取图像的局部特征;然后,计算图像的全局颜色特征和信息熵;最后,利用信息熵动态分配全局颜色特征和局部特征的权重,计算图像间的相似度进行图像检索。实验结果表明:该方法的检索性能优于颜色直方图法和尺度不变特征变换算法。
In view of low precision rate and recall rate problems on image retrieval based on single global color feature or local feature,an image retrieval method based on information entropy and improved scale-invariant feature transform(SIFT) algorithm was proposed. First,local feature of images was extracted by an improved SIFT algorithm. Then,color feature and information entropy were calculated. Finally,the weights of color feature and SIFT feature based on information entropy were utilized to calculate the similarity among images for image retrieval. The experimental results show that this method has better retrieval performance than the color histogram and SIFT algorithm.
出处
《河南科技大学学报（自然科学版）》
CAS
北大核心
2014年第6期42-46,7,共5页
Journal of Henan University of Science And Technology:Natural Science
基金
海南省科技兴海专项基金项目(XH201311)
