An Efficient Image Indexing Method Based on Class Specific Hyper Graph

Abstract:

An efficient image indexing method based on class specific hyper graph is proposed. The presented indexing method works efficiently and the relevance of the original image data is enhanced. Because of that an ordered image database benefits the efficient searching. The relevance of images depends on the similarity between different images. According to clustering theory, we can take any sample image in the database as one clustering center, and then the siblings of the center and their siblings are consistently searched, which is known as similarity spread. After that, the disordered image database is sorted out and the searching result is not tedious any more. The proposed method has been tested by an open arts press image database, which shows that our method can obviously improve the indexing speed. Moreover, the indexing results make the whole image searching system capable of association.

• Periodical: Applied Mechanics and Materials (Volumes 513-517)
• Pages: 2914-2918
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.513-517.2914
• Online since: February 2014
• Authors: Kun Lu*, Sheng Ping Xia, Song Hua He
• Keywords: CBIR, CSHG, RSOM, SIFT, Similarity Spread

References

[1] Office Clip Art Images, http: /office. microsoft. com.

[2] http: /www. flickr. com.

[3] http: /www. fotolia. com.

[4] He Ling, Wu Lingda, Cai Yichao. Survey of Indexing Technology in CBIR[J]. Mini and Micro Computer System, 2006. Vol. 27, Issue 1: pp.141-145.

[5] Clarlson, K., Nearest-neighbor searching and metric space dimensions", In Nearest-Neighor Methods for Learning and Vision: Theory and Practice. (MIT Press, 2005).

DOI: 10.7551/mitpress/4908.003.0005

[6] Liu Jianjun. CHSG Reconstruction and Object Recognition Based on Image Local Invariant Feature[D]. National University of Defense Technology. (2009).

[7] David G. Lowe. Object Recognition from Local Scale-Invariant Feature, Proc. of the International Conference Computer Vision, Corfu(Sept. 1999).

[8] K. Mikolajczyk and C. Schmid, Indexing Based on Scale Invariant Interest Points, " Proc. Eighth Int, l Conf. Computer Vision, pp.525-531, (2001).

DOI: 10.1109/iccv.2001.937561

[9] C. Schmid and R. Mohr, Local Grayvalue Invariants for Image Retrieval, IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 19, no. 5, pp.530-534, May (1997).

DOI: 10.1109/34.589215

[10] Xia Shengping，Liu Jianjun，Yuan zhengtao，Yu Hua，Zhang Lefeng，Yu Wenxian. Clustering Based on Clusters of Paralleled Distributed RSOM[J]. ElectronicJournal，2007，35(3): 385-391.

[11] Xia Shengping, Zhang Lefeng, Yu Hua, Zhang Jing, Hu WeiDong, Yu WenXian. The Research of Theory and Algorithm about Machine Learning Based on the Tree Model of RSOM. ACTA Electronic SINICA, Vol. 33 No. 5.

[12] Zheng Junjun, Xia Shengping, Li Xinguang, Zhu Yiwei, Liu JianJun, Tan Liqiu. K Nearest Neighbors Detecting Algorithm Based on a RSOM Tree. Journal of Shandong University (Engineering Science ), Vol 41 No2 1672-3961(2011)02-0080-05.