Optical correlation-based cross-domain image retrieval system.

Researchers

Journal

Modalities

Models

Abstract

A novel cross-domain image retrieval system that is based on a high-speed optical correlator with a coaxial holographic system is presented. Our newly designed conversion module for the optical correlator allows various kinds of data to be converted to pagedata with a uniform optical intensity by using an autoencoder, which is difficult with other conventional methods. By using our conversion module, an existing model for deep learning could be utilized as a feature extractor. A sketch-based cross-domain image retrieval system with the goal of discovering similar photos by querying freehand human sketches was experimentally demonstrated using our optical correlator. We believe that this proposed optical correlation-based system helps expand the applications of the optical correlator.