應用於拍賣網頁分類目錄下的影像搜尋系統

A Category-Based Image Retrieval System

Applied to Online Auction

指導老師:劉震昌博士  研究生:周怡君  日期:中華民國 723

 

論文摘要

近年來,隨著多媒體資訊的進步,數位相機、具有照像功能的手機等相關科技產品的普及化,許多研究學者開始採用影像的內容來做為圖片搜尋的依據。本論文主要是將此方法應用在Yahoo拍賣網頁上,並且在特定的分類目錄下,實做一個影像搜尋系統。

本搜尋系統,提供使用者上傳或點選資料庫中影像做搜尋。傳統影像搜尋的方法使用整張圖的影像特徵,本論文提出了一種在影像中擷取物件區域的方法,利用邊緣偵測找出物件邊緣,再利用 Principle Component Analysis 計算邊緣點聚集的區域,此區域即可能是影像中重要的物件區域。本論文利用Yahoo拍賣網頁抓取的資料以驗證本系統的可行性,實驗証實此種區域偵測的方法有70% 的準確率,並且利用區域的特徵進行搜尋,在搜尋的 R-Precision較利用整張影像特徵搜尋有顯著的改進。本論文亦實作了文字搜尋與文字及影像合併搜尋,與上述影像搜尋比較其效能

 

關鍵詞:影像內容搜尋系統、色彩直方圖、區域擷取、PCA

簡介

CBIR主要流程(Fig 1)1.由使用者輸入影像;2.根據影像擷取特徵,3.最後做相似度的比對。而其最主要的重點在於,如何從影像中擷取出重要的區域、擷取哪些特徵、特徵如何做比對與排序問題,其中如何從影像中找到重要的區域最為困難,也是本論文中最主要探討的地方。

Fig 1

一般拍賣網頁的使用者,通常針對影像中的物件來做為搜尋的主要目的,所以如何從影像中找出物件是很重要的關鍵;因此,在本論文中我們提出一個從影像中切割出物件區域影像的方法以提高搜尋的準確率(Fig 2由左至右:原圖,經過Sobel filter轉換成包含特徵的圖片,最後圈出橘色區域)

Fig 2

實驗結果

    Fig 3為我們挑選的Query Example,比較整張影像搜尋的結果(Fig 5Fig 6上排)與我們提出的方法-在影像中找到一個具代表性的區域-搜尋結果(Fig 5Fig 6 下排)利用R-Precision來評估(Fig 4)。由Fig 5Fig 6對照,可以很明顯的發現,我們提出的方法搜尋結果比整張影像搜尋結果正確,再根據Fig 4R-Precision,準確率確實是提升了。(Fig 5:以圖找圖結果、Fig 6:以圖找網頁結果,其中的紅色框框為使用者想得到的回傳結果)

 

GT4.png

Fig 3

Fig 4

R5.png

Fig 5

R6.png

Fig 6

文獻參考

[1] M. Flickner, H. Sawhney, W. Niblack, J. Ashley, Q. Huang, B. Dom, M. Gorkani, J. Hafner, D. Lee, D. Petkovic, D. Steele, and P. Yanker, "Query by image and video content: The QBIC system." IEEE Computer, vol. 28, pp. 23-32, Sept, 1995.

[2] W. Niblack, R. Barber, W. Equitz, M. Flickner, E. Glasman, D. Petkovic, P. Yanker, C. Faloutsos and G. Taubin, "The QBIC project: Querying images by content using colour, texture and shape, " in Proceedings of the SPIE on Storage and Retrieval for Image and Video Databases, vol. 1908, pp. 173-187, 1993.

[3] http://www.hermitagemuseum.org/fcgi-bin/db2www/qbicSearch.mac/qbic?selLang=E

nglish

[4] M. J. Swain and D. H. Ballard, "Color indexing," International Journal of Computer Vision, vol 7, pp. 11-32, 1991.

[5] B. S. Manjunath, and W. Y. Ma, "Texture features for browsing and retrieval of image data," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 18, pp. 837-842, Aug, 1996.

[6] S. Newsam, B. Sumengen, and B. S. Manjunath, "Category-based image retrieval," on Proceedings of the 2001 International Conference on Image Processing, vol. 3, pp. 596-599, 2001.

[7] M. N. Do, and M. Vetterli, "Wavelet-Based Texture Retreival Using Generalized Gaussian Density and Kullback-Leibler Distance," IEEE Transactions on image processing, vol. 11, Feb, 2002.

[8] B. G. Prasad, K. K. Biswas, and S. K. Gupta, "Region-Based Image Retrieval using Integrated color, shape, and location index," Computer vision and image Understanding, vol. 94, pp. 193-233, 2004.

[9] A. k. Jain and A. Vailaya, "Image Retrieval using Color and Shape," Pattern Recognition, vol. 29, pp.1233-1244, 1996.

[10] J. Han, M. Kamber, Data Mining: Concepts and Techniques, pp. 340-341, 2001.

[11] Y. Rubner, C. Tomasi and L. J. Guibas, "The Earth Mover’s Distance as a Metric for Image Retrieval," International Journal of Computer Vision, vol. 40, pp. 99-121, 2000

[12] C. Carson, S. Belongie, H. Greenspan and J. Malik, "Blobworld: Image segmentation using Expectation-Maximization and its application to image querying," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 24, pp. 1026-1038, Aug, 2002.

[13] Y. A. Aslandogan and C. T. Yu, "Evaluating strategies and systems for content based indexing of person images on the web," on ACM Multimedia, pp.313-321, 2000.

[14] A. R. Smith, "Color gamut transform pairs," on ACM SIGGRAPH Computer Grapics, vol 12, pp. 12-19, 1978

[15] J. R. Parker, John Wiley & Sons, Algorithms for Image Processing and Computer Vision, pp. 16-19, 1997.

[16] M. Sonka, V. Hlavac, R. Boyle, Image Processing, Analysis, and Machine Vision, 1998.