| | |
Summary: Learning Embeddings for Fast Approximate Nearest
Neighbor Retrieval
Vassilis Athitsos, Jonathan Alon, Stan Sclaroff, and George Kollios #
Computer Science Department
Boston University
111 Cummington Street
Boston, MA 02215
email: {athitsos, jalon, sclaroff, gkollios}@cs.bu.edu
Abstract
We present an embedding method that can significantly reduce nearest neighbor re
trieval time when the underlying distance measure is computationally expensive. Database
and query objects are embedded into a Euclidean space, in which similarities can be
rapidly measured using a weighted Manhattan distance. Embedding construction is
formulated as a machine learning task, where AdaBoost is used to combine many sim
ple, 1D embeddings into a multidimensional embedding that preserves a significant
amount of the proximity structure in the original space. Performance is evaluated in
a hand pose estimation system, and a dynamic gesture recognition system, where the
proposed method is used to retrieve approximate nearest neighbors under expensive
similarity measures. In both systems, BoostMap significantly increases efficiency, with
minimal losses in accuracy. Moreover, the experiments indicate that BoostMap com
|
