目前基于单一特征的图像检索方法已经成熟, 如何有效地利用这些特征是图像检索的一个关键问题。利用颜色、纹理、形状等特征进行图像检索, 结合自反馈技术中的权系数调整方法, 将动态生成权值的方法应用于多特征融合过程以实现检索。实验结果表明算法能够准确、高效地检索出目标图像。相对于单一特征的检索, 本文方法有效地提高了图像检索的精度。
Image retrieval based on single feature has already been quite mature.How to use these single features effectively is the key to the image retrieval.In this paper, we apply dynamic weight to multi-featured process to achieve retrieval by combining the single features such as color, texture and shape with the feedback technology.Experimental results show that the algorithm using dynamic weight can accurately and efficiently retrieve the target image.
[1]孙君顶, 原芳.基于内容的图像检索技术[J].计算机系统应用, 2011, 20(8):240-244.
[2]马凌蛟.基于颜色和形状特征的图像检索技术及其应用[D].长春:吉林大学, 2011.
[3]Hideyuki Tamura, Shun ji Mori, Takashi Yamawaki.Texture features corresponding to visual perception [J].IEEE Trans.On Sys, Man, and Cyb, 1978, SMC-8(6):460-473.
[4]Stricker M, Orengo M.Similarity of color images[J].Proc.SPIE Storage and Retrieval for Image and Video Database.San Jose CA USA, 1995, 2420pp:381-392.
[5]Pass G, Zabih R, Miller J.Comparing images using color coherence vectors[C].4th ACM Conf.on Multimedia.Boston, ACM, 1996.
[6]J.Huang, S.R.Kumar, et al.Image Indexing using color correlograms [C].IEEE Conference on Computer Vision and Pattern Recognition, 1997:762-768.
[7]曾俊.图像边缘检测技术及其应用研究[D].武汉:华中科技大学, 2011.
[8]Zhang D S, Lu G.Generic Fourier Descriptors for Shape-based Image Retrieval [J].In Proc of IEEE International Conference on Multimedia and Expo, Lausanne, Switzerland, 2002.