Using Apache Lucene to Search Vector of Locally Aggregated Descriptors

TL;DR

This paper extends Surrogate Text Representation (STR) to efficiently search VLAD vectors using Apache Lucene, eliminating the need for reordering results and maintaining high accuracy in visual similarity search.

Contribution

The paper introduces a novel extension of STR tailored for VLAD vectors, enabling efficient similarity search without reordering, improving upon baseline STR performance.

Findings

Extended STR outperforms baseline STR in experiments

Achieves near-original VLAD vector performance

Eliminates reordering phase in similarity search

Abstract

Surrogate Text Representation (STR) is a profitable solution to efficient similarity search on metric space using conventional text search engines, such as Apache Lucene. This technique is based on comparing the permutations of some reference objects in place of the original metric distance. However, the Achilles heel of STR approach is the need to reorder the result set of the search according to the metric distance. This forces to use a support database to store the original objects, which requires efficient random I/O on a fast secondary memory (such as flash-based storages). In this paper, we propose to extend the Surrogate Text Representation to specifically address a class of visual metric objects known as Vector of Locally Aggregated Descriptors (VLAD). This approach is based on representing the individual sub-vectors forming the VLAD vector with the STR, providing a finer representation of the vector and enabling us to get rid of the reordering phase. The experiments on a publicly available dataset show that the extended STR outperforms the baseline STR achieving satisfactory performance near to the one obtained with the original VLAD vectors.