譚志偉++孫新領(lǐng)++孫挺
摘 要: 針對(duì)高級(jí)用戶的描述對(duì)象與低級(jí)圖像特征之間的語義差異問題,提出一種基于自組織特征重加權(quán)和相關(guān)反饋的CBIR算法。首先對(duì)查詢圖像和數(shù)據(jù)庫圖像采用Gabor小波變換和小波矩技術(shù)提取圖像特征向量;然后進(jìn)行相似性度量,同時(shí)為了最大程度地從相關(guān)圖像中分離非相關(guān)圖像,引入自組織特征重加權(quán)模式,確保非相關(guān)圖像集沒有單一的相關(guān)圖像;最后將用戶反饋和特征加權(quán)循環(huán)進(jìn)行,直到得出用戶滿意的結(jié)果。仿真實(shí)驗(yàn)在Corel收集的1 000幅圖像庫上進(jìn)行,對(duì)某些類別的圖像,該算法的檢索精度可高達(dá)97.5%,在無噪聲情況下,對(duì)于前10幅圖像,該算法的準(zhǔn)確率為82.78%,對(duì)于前100幅圖像,精度僅降到66.70%,在有噪聲情況下,精度下降僅3%左右。相比其他優(yōu)秀算法,該算法具有更高的精度和更好的噪聲魯棒性。
關(guān)鍵詞: 圖像特征; 基于內(nèi)容的圖像檢索; 自組織特征重加權(quán); Gabor小波變換; 小波矩
中圖分類號(hào): TN911.73?34; TP391.4 文獻(xiàn)標(biāo)識(shí)碼: A 文章編號(hào): 1004?373X(2016)23?0047?05
CBIR algorithm based on relevance feedback technology
and self?organized feature reweighting
TAN Zhiwei1, SUN Xinling1, SUN Ting2
(1. Department of Computer Science & Technology, Henan Institute of Technology, Xinxiang 453003, China;
2. Institute of Visualization, Northwest University, Xian 710069, China)
Abstract: To solve the problem of semantic difference between the description object of the advanced user and low?level image feature, a content?based image retrieval (CBIR) algorithm based on relevance feedback (RF) technology and self?organized feature reweighting is proposed. The Gabor wavelet transform and wavelet moment technology are used to extract the image feature vectors of queried image and database image, and then the similarity is measured. In order to separate the non?relevance image from the relevance image to the maximum extent, the self?organized feature reweighting mode is introduced to ensure there is no any single relevance image in the non?relevance image set. The user feedback and feature weighting are conducted circularly until the user obtains a satisfactory result. The simulation experiments are performed on 1 000 images collected by Corel. The retrieval accuracy of the algorithm can reach up to 97.5% for some certain images. Under the condition of no noise, the algorithm accuracy for first 10 images can reach up to 82.78%, and the accuracy for first 100 images is reduced only to 66.70%. The accuracy under the noise condition is decreased by 3%. In comparison with other outstanding algorithms, this algorithm has higher accuracy and better noise robustness.
Keywords: image feature; content?based image retrieval; self?organized feature reweighting; Gabor wavelet transform; wavelet moment
0 引 言
基于內(nèi)容的圖像檢索(Content?based Image Retrie?val,CBIR)是機(jī)器視覺和模式識(shí)別領(lǐng)域最熱門的研究課題之一[1?2]。在CBIR系統(tǒng)中,視覺特征的提取依賴于特征的表征方式,與相似性匹配直接關(guān)聯(lián),而相似性評(píng)估常常會(huì)導(dǎo)致語義差[3?4],即高級(jí)用戶描述的對(duì)象與低級(jí)圖像特征之間的語義差異,同時(shí)也考慮查詢圖像自身的影響,如有噪聲污染等。因此,減少語義差異和提高特征提取的效果是CBIR的主要目標(biāo)[5]。
已有很多研究者對(duì)CBIR進(jìn)行了研究,例如,文獻(xiàn)[6]提出基于灰度級(jí)共生矩陣CBIR算法,灰度共生矩陣既可以表征亮度分布,又可以表征亮度與像素位置關(guān)系,因此,廣泛應(yīng)用于紋理特征。然而,在復(fù)雜紋理情況下表現(xiàn)并不穩(wěn)定。
文獻(xiàn)[7]將粒子群優(yōu)化算法的進(jìn)化搜索過程與用戶的反饋過程有效結(jié)合,提出了一種基于粒子群的CBIR方法,本文簡(jiǎn)稱為SO?RF算法,避免了初始檢索對(duì)用戶認(rèn)知的影響以及對(duì)反饋效果造成的局限性,然而,很容易出現(xiàn)相關(guān)反饋提供的標(biāo)注樣本數(shù)不足的問題。
文獻(xiàn)[8]使用智能搜索進(jìn)行特征選擇,利用改進(jìn)的二進(jìn)制引力搜索(IBGSA)選擇最佳功能子集,改善分類的精度。雖然提高了分類精度,但I(xiàn)BGSA的復(fù)雜度比二進(jìn)制引力搜索(BGSA)的復(fù)雜性高了很多。
文獻(xiàn)[9]通過合并未標(biāo)記的圖像,提出了一種基于改進(jìn)型相關(guān)反饋的CBIR算法,通過使用已標(biāo)記數(shù)據(jù)觀測(cè)誤差函數(shù)最小化,選擇綜合性能最好的回歸函數(shù),同時(shí)兼顧圖像的語義特征及圖像空間的幾何結(jié)構(gòu),然而,算法依然存在視覺特征和語義信息中的“語義鴻溝”問題。
為了提高分類精度,本文在相關(guān)反饋的基礎(chǔ)上,采用了自組織特征重加權(quán)方法,同時(shí)在特征向量和特征表征中運(yùn)用了Gabor小波和小波矩技術(shù),稱之為SOFR?RF算法,實(shí)驗(yàn)結(jié)果表明,本文算法具有更高的精度和更好的魯棒性,甚至在有噪聲的情況下也是如此。
3 結(jié) 論
本文提出了一種基于自組織特征重加權(quán)和相關(guān)反饋的CBIR算法,利用自組織特征重加權(quán)模式盡可能地從相關(guān)圖像中分離非相關(guān)圖像,確保非相關(guān)圖像集沒有單一的相關(guān)圖像,圖像檢索的用戶作用通過相關(guān)反饋體現(xiàn)。在有噪聲和無噪聲情況下的實(shí)驗(yàn)結(jié)果表明,本文算法具有較高的檢索準(zhǔn)確度和較好的魯棒性。
參考文獻(xiàn)
[1] BUGATTI P H, KASTER D S, PONCIANO?SILVA M, et al. PRoSPer: perceptual similarity queries in medical CBIR systems through user profiles [J]. Computers in biology & medicine, 2014, 45(2): 8?19.
[2] 王明三.基于內(nèi)容的圖像檢索關(guān)鍵技術(shù)研究[D].成都:電子科技大學(xué),2014.
[3] 施智平,李清勇,趙曉東,等.基于內(nèi)容圖像檢索中的優(yōu)化鑒別特征[J].計(jì)算機(jī)輔助設(shè)計(jì)與圖形學(xué)學(xué)報(bào),2012,24(12):1592?1598.
[4] QUELLEC G, LAMARD M, CAZUGUEL G, et al. Fast wavelet?based image characterization for highly adaptive image retrieval [J]. IEEE transactions on image processing, 2011, 21(4): 1613?1623.
[5] 李宗民,唐志輝.九宮格空間框架的圖書圖像檢索[J].中國圖象圖形學(xué)報(bào),2013,18(3):325?329.
[6] 于艷紅.基于顏色和紋理特征的圖像檢索與應(yīng)用[D].大連:大連理工大學(xué),2013.
[7] 許相莉,張利彪,劉向東,等.基于粒子群的圖像檢索相關(guān)反饋算法[J].電子學(xué)報(bào),2010,38(8):1935?1940.
[8] RASHEDI E, NEZAMABADI H. Feature subset selection using improved binary gravitational search algorithm [J]. Journal of intelligent and fuzzy systems, 2014, 26(3): 1211?1221.
[9] LI G. Improving relevance feedback in image retrieval by incorporating unlabelled images [J]. Indonesian journal of electrical engineering, 2013, 11(7): 3634?3640.
[10] PIRAS L, GIACINTO G. Neighborhood?based feature weighting for relevance feedback in content?based retrieval [C]// Procee?dings of 2009 Image International Workshop on Analysis for Multimedia Interactive Services. Italy: IEEE, 2009: 238?241.
[11] WANG J Z. Corel database [EB/OL]. [2006?03?23]. http://wang.ist.psu.edu/docs/related.html.
[12] 李宗民,唐志輝.九宮格空間框架的圖書圖像檢索[J].中國圖象圖形學(xué)報(bào),2013,18(3):325?329.
[13] 李波.正交圖像矩及其在精確定位與產(chǎn)品檢測(cè)中的應(yīng)用研究[D].武漢:華中科技大學(xué),2012.
[14] ALVAREZ M J, ELENA G, FRANCESCO B, et al. Color and texture features for retrieval in granite industry [J]. Dyna. Rev. Fac. Nac. Minas., 2010, 77(161): 121?130.
[15] JIAN M, GUO H, LIU L. Texture image classification using visual perceptual texture features and Gabor wavelet features [J]. Journal of computers, 2009, 4(8): 763?770.
[16] 李世雄.小波變換及其應(yīng)用[J].高等數(shù)學(xué)研究,2002,5(1):43?45.