Large Scale Deep Convolutional Neural Network Features Search with Lucene

TL;DR

This paper presents a scalable method for indexing deep CNN features as text using Lucene, enabling efficient content-based image retrieval on massive datasets of around 100 million images.

Contribution

The authors introduce a novel approach to convert CNN features into textual form for indexing with Lucene, supporting large-scale image retrieval combining text and visual search.

Findings

Effective indexing of CNN features as text for large datasets

Optimized textual representation improves index size and response time

Prototype demonstrates retrieval on 100 million images

Abstract

In this work, we propose an approach to index Deep Convolutional Neural Network Features to support efficient content-based retrieval on large image databases. To this aim, we have converted the these features into a textual form, to index them into an inverted index by means of Lucene. In this way, we were able to set up a robust retrieval system that combines full-text search with content-based image retrieval capabilities. We evaluated different strategies of textual representation in order to optimize the index occupation and the query response time. In order to show that our approach is able to handle large datasets, we have developed a web-based prototype that provides an interface for combined textual and visual searching into a dataset of about 100 million of images.