IDEAL: Independent Domain Embedding Augmentation Learning

TL;DR

The paper introduces IDEAL, a novel data transformation technique that learns multiple independent embedding spaces for different data domains, significantly improving visual retrieval performance when combined with existing deep metric learning methods.

Contribution

IDEAL is a new mechanism that learns multiple independent embedding spaces for data transformations, enhancing deep metric learning without altering existing loss functions.

Findings

IDEAL improves MS loss performance on Cars-196 from 84.5% to 87.1%.

IDEAL achieves state-of-the-art results on Cars-196, CUB-200, and SOP benchmarks.

IDEAL outperforms recent DML methods like Circle loss and XBM.

Abstract

Many efforts have been devoted to designing sampling, mining, and weighting strategies in high-level deep metric learning (DML) loss objectives. However, little attention has been paid to low-level but essential data transformation. In this paper, we develop a novel mechanism, the independent domain embedding augmentation learning ({IDEAL}) method. It can simultaneously learn multiple independent embedding spaces for multiple domains generated by predefined data transformations. Our IDEAL is orthogonal to existing DML techniques and can be seamlessly combined with prior DML approaches for enhanced performance. Empirical results on visual retrieval tasks demonstrate the superiority of the proposed method. For example, the IDEAL improves the performance of MS loss by a large margin, 84.5\% $\rightarrow$ 87.1\% on Cars-196, and 65.8\% $\rightarrow$ 69.5\% on CUB-200 at Recall$@1$. Our IDEAL with MS loss also achieves the new state-of-the-art performance on three image retrieval benchmarks, \ie, \emph{Cars-196}, \emph{CUB-200}, and \emph{SOP}. It outperforms the most recent DML approaches, such as Circle loss and XBM, significantly. The source code and pre-trained models of our method will be available at\emph{\url{https://github.com/emdata-ailab/IDEAL}}.