A Dense-Depth Representation for VLAD descriptors in Content-Based Image Retrieval

TL;DR

This paper introduces a dense-depth representation method for VLAD descriptors using CNN feature maps, enhancing image retrieval performance by increasing feature density and tested on multiple public datasets.

Contribution

It proposes a new detector on CNN feature maps to generate more features, improving VLAD-based image retrieval accuracy.

Findings

Improved retrieval performance on Holidays, Oxford5k, Paris6k, and UKB datasets.

Enhanced feature density leads to better aggregation in VLAD descriptors.

Method outperforms existing approaches in benchmark tests.

Abstract

The recent advances brought by deep learning allowed to improve the performance in image retrieval tasks. Through the many convolutional layers, available in a Convolutional Neural Network (CNN), it is possible to obtain a hierarchy of features from the evaluated image. At every step, the patches extracted are smaller than the previous levels and more representative. Following this idea, this paper introduces a new detector applied on the feature maps extracted from pre-trained CNN. Specifically, this approach lets to increase the number of features in order to increase the performance of the aggregation algorithms like the most famous and used VLAD embedding. The proposed approach is tested on different public datasets: Holidays, Oxford5k, Paris6k and UKB.