Image Data Compression for Covariance and Histogram Descriptors

TL;DR

This paper introduces two data compression methods for covariance and histogram image descriptors that significantly reduce dataset size while maintaining or improving classification accuracy, enabling faster and more practical image analysis.

Contribution

The paper presents novel supervised compression techniques for covariance and histogram descriptors that drastically reduce dataset size with minimal impact on accuracy.

Findings

Data sets can be reduced to 16% or as low as 2% of original size.

Compressed datasets can match or outperform full dataset classification accuracy.

Compression drastically reduces test-time computation.

Abstract

Covariance and histogram image descriptors provide an effective way to capture information about images. Both excel when used in combination with special purpose distance metrics. For covariance descriptors these metrics measure the distance along the non-Euclidean Riemannian manifold of symmetric positive definite matrices. For histogram descriptors the Earth Mover's distance measures the optimal transport between two histograms. Although more precise, these distance metrics are very expensive to compute, making them impractical in many applications, even for data sets of only a few thousand examples. In this paper we present two methods to compress the size of covariance and histogram datasets with only marginal increases in test error for k-nearest neighbor classification. Specifically, we show that we can reduce data sets to 16% and in some cases as little as 2% of their original size, while approximately matching the test error of kNN classification on the full training set. In fact, because the compressed set is learned in a supervised fashion, it sometimes even outperforms the full data set, while requiring only a fraction of the space and drastically reducing test-time computation.