圖像檢索系統(tǒng)外文翻譯-資料下載頁(yè)
【導(dǎo)讀】我們提出基于文本和內(nèi)容相結(jié)合的圖像檢索系統(tǒng)。我們的想法是在原來(lái)標(biāo)。這個(gè)系統(tǒng)包含兩種不同的用戶界面：終端用戶必須在Flickr網(wǎng)站。搜索特征定義距離函數(shù)和權(quán)值，以得到更好的搜索效果。搜索提供視覺(jué)相似度較近的搜索結(jié)果。由于兩種方法的獨(dú)立性，很可能它們的結(jié)。合能夠提高搜索系統(tǒng)的性能，依靠?jī)煞N方法的益處。目前圖像搜索系統(tǒng)已經(jīng)實(shí)現(xiàn)用Java，C++和PostgreSQL完成。象的集合使用SAPIR[3]從Flickr網(wǎng)站分離出來(lái)。在離線階段，使用SAPIR集合。提供的URL地址把圖片從Flickr網(wǎng)站下載下來(lái)。為每幅圖像建立六種特征向量，三個(gè)使用。詞干提取算法，三個(gè)不使用。最大距離做的歸一化操作之后執(zhí)行一個(gè)k-NN搜索。最后，搜索結(jié)果會(huì)顯示給用戶，并且同。索，同時(shí)這個(gè)結(jié)果圖像的鏈接將發(fā)布到Flickr站點(diǎn)上。時(shí)間主要依賴(lài)于圖片進(jìn)行比較的特征數(shù)量和所使用的距離函數(shù)。采用的樣本圖片在Flickr站點(diǎn)超過(guò)115,000張圖像，從薩皮爾集合獲得。這項(xiàng)工作部分被(智利)11070037FONDECYT項(xiàng)目支持。媒體信息檢索國(guó)際研討會(huì)的學(xué)報(bào)，137~146頁(yè).美國(guó)計(jì)算機(jī)協(xié)會(huì)出版,2020.
【正文】 system is installed on a desktop puter with an AMD Athlon XP 1800+ and 1 GB RAM running CentOS with Sun Java , Jakarta Tomcat and PostgreSQL installed. The memory usage of the VM is about 90 MB (this includes the memory used by Tomcat and our system). When the user does not enter any search text and the query object does not contain textbased features, a linear search over all the collection is performed. In this case, the search time depends mainly on the number of pared features and the distance functions used. Using all the features available in the system, the search time of the linear scan is about 20 seconds. Using only the RGB Histogram 333 and Euclidean distance, the search time of a linear scan is about 1 second. It is worth noting that the images features are retrieved from a file during the search and are not stored in main memory. Figure 2. The advanced search. When the user enters any search text or the query contains textbased features, the inverted file is used for retrieving the set of images that contain them. In this case, on average, the search time is reduced to less than a second. VI. CONCLUSIONS In this work, we presented a system that allows a bined search of textbased and contentbased features on a sample of more than 115,000 images of Flickr obtained from the SAPIR collection. We are evaluating the inclusion of the M3tree [2] into the system as an indexing structure for dynamic binations of metric distances. ACKNOWLEDGMENT This work was partially supported by FONDECYT (Chile) Project 11070037. REFERENCES [1] R. A. BaezaYates and B. RibeiroNeto. “Modern Information Retrieval”. AddisonWesley Longman Publishing Co., Inc., Boston, MA, USA, 1999. [2] B. Bustos and T. Skopal. “Dynamic similarity search in multimetric spaces”. In Proc. 8th ACM SIGMM International Workshop on Multimedia Information Retrieval (MIR’06), pages 137–146. ACM Press, 2020. [3] F. Falchi, M. Kacimi, Y. Mass, F. Rabitti, and P. Zezula. “SAPIR: Scalable and distributed image searching”. In SAMT (Posters and Demos), vol. 300 of CEUR Workshop Proceedings, pages 11–12, Genoa, Italy, November 2020. [4] M. F. Porter. “An algorithm for suffix stripping”. Program, 14(3):130137, July 1980. martin/PorterStemmer/
鄂ICP備17016276號(hào)-1