An Adaptive Index Structure for High-Dimensional Similarity Search

A practical method for creating a high dimensional index structure that adapts to the data distribution and scales well with the database size, is presented. Typical media descriptors, such as texture features, are high dimensional and are not uniformly distributed in the feature space. The performance of many existing methods degrade if the data is not uniformly distributed. The proposed method offers an efficient solution to this problem. First, the data’s marginal distribution along each dimension is characterized using a Gaussian mixture model. The parameters of this model are estimated using the well known ExpectationMaximization (EM) method. These model parameters can also be estimated sequentially for on-line updating. Using the marginal distribution information, each of the data dimensions can be partitioned such that each bin contains approximately an equal number of objects. Experimental results on a real image texture data set are presented. Comparisons with existing tech...

• Type: Conference
• Year: 2001
• Where: PCM
• Authors: Peng Wu, B. S. Manjunath, Shivkumar Chandrasekaran
• Keywords: Data’s Marginal Distribution, Marginal Distribution, Multimedia, PCM 2001, Typical Media Descriptors