Fast binary embedding via circulant downsampled matrix
Journal
Proceedings - International Conference on Image Processing, ICIP
Journal Volume
2016-August
Pages
1789-1793
Date Issued
2016
Author(s)
Abstract
Binary embedding of high-dimensional data aims to produce low-dimensional binary codes while preserving discriminative power. State-of-the-art methods often suffer from high computation and storage costs. We present a simple and fast embedding scheme by first downsampling N-dimensional data into M-dimensional data and then multiplying the data with an M×M circulant matrix. Our method requires O(N + M log M) computation and O(N) storage costs. We prove if data have sparsity, our scheme can achieve similarity-preserving well. Experiments further demonstrate that though our method is cost-effective and fast, it still achieves comparable performance in image applications. © 2016 IEEE.
Subjects
Circulant matrix; Dimensionality reduction; Embedding; Random projection; Subsampling
Other Subjects
Bins; Clustering algorithms; Cost effectiveness; Costs; Digital storage; Image processing; Circulant matrix; Dimensionality reduction; Embedding; Random projections; Subsampling; Matrix algebra
Type
conference paper