Deep Image Retrieval is not Robust to Label Noise

TL;DR

This paper demonstrates that deep image retrieval models are more sensitive to label noise than classification models, and it explores various types of label noise specific to image retrieval to assess their impact.

Contribution

It is the first study to analyze the effects of label noise on deep image retrieval, including different noise types and their influence on model robustness.

Findings

Image retrieval models are less robust to label noise than classification models

Different types of label noise uniquely affect retrieval performance

Understanding noise effects can guide better dataset curation for retrieval tasks

Abstract

Large-scale datasets are essential for the success of deep learning in image retrieval. However, manual assessment errors and semi-supervised annotation techniques can lead to label noise even in popular datasets. As previous works primarily studied annotation quality in image classification tasks, it is still unclear how label noise affects deep learning approaches to image retrieval. In this work, we show that image retrieval methods are less robust to label noise than image classification ones. Furthermore, we, for the first time, investigate different types of label noise specific to image retrieval tasks and study their effect on model performance.