Abstract—In the recent years, lots of research on image databases has led to proposition of different kinds of multi-dimensional indexing structures to support similarity search in image database systems. Most of current commercial multi-dimensional image indexing structures cannot manage image data points in high dimensional spaces. Such that sequential scanning of these high-dimensional data points can be faster than using a multi-dimensional indexing structure. Therefore, another class of multi-dimensional indexing structures especially has developed for high-dimensional spaces. Due to importance and variety of high-dimensional indexing structures, we first classified them based on their main idea and then we evaluated them according to suggested design requirements of desired high-dimensional indexing structures. We hope this proposed framework will lead to development of more efficient structures to support similarity search in high-dimensional spaces.
Index Terms—High-dimensional space, multi-dimensional indexing structure, image database.
Mohammadreza keyvanpour is with Department of Computer Engineering, Alzahra University, Tehran, Iran (e-mail: Keyvanpour@alzahra.ac.ir).
Najva izadpanah izadpanah was with Department of Computer Engineering, Islamic Azad University,Qazvin branch, Qazvin, Iran(e-mail: n.izadpanah@qiau.ac.ir)
Haleh karbasforoushan was with Department of Computer Science, University of Southern California, Los Angeles, CA, USA.
Cite: Mohammadreza Keyvanpour, Najva Izadpanah, and Haleh Karbasforoushan, "Classification and Evaluation of High-dimensional Image Indexing Structures," International Journal of Machine Learning and Computing vol. 2, no. 3, pp.252-256, 2012.