Deep Image Retrieval is not Robust to Label Noise

  title={Deep Image Retrieval is not Robust to Label Noise},
  author={Stanislav Dereka and I. A. Karpukhin and Sergey Kolesnikov},
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… 

Figures and Tables from this paper


Learning with Label Noise for Image Retrieval by Selecting Interactions
A noise-resistant method for image retrieval named Teacher-based Selection of Interactions, T-SINT 1, which identifies noisy interactions, i.e. elements in the distance matrix, and selects correct positive and negative interactions to be considered in the retrieval loss by using a teacher-based training setup which contributes to the stability.
Efficient Two-stage Label Noise Reduction for Retrieval-based Tasks
A noisy label reduction method for text datasets is proposed, which could be applied at retrieval-based tasks by getting a conditional independent probability distribution to identify noisy labels accurately and can improve the label noise reduction accuracy and end-to-end classification accuracy.
Image Classification with Deep Learning in the Presence of Noisy Labels: A Survey
Leveraging Weakly Annotated Data for Fashion Image Retrieval and Label Prediction
This paper presents a method to learn a visual representation adapted for e-commerce products based on weakly supervised learning that achieves nearly state-of-art results on the DeepFashion In-Shop Clothes Retrieval and Categories Attributes Prediction tasks.
Probabilistic End-To-End Noise Correction for Learning With Noisy Labels
  • Kun Yi, Jianxin Wu
  • Computer Science
    2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
  • 2019
An end-to-end framework called PENCIL, which can update both network parameters and label estimations as label distributions and is more general and robust than existing methods and is easy to apply.
Learning Visual Features from Large Weakly Supervised Data
This paper trains convolutional networks on a dataset of 100 million Flickr photos and comments, and shows that these networks produce features that perform well in a range of vision problems.
An Entropic Optimal Transport Loss for Learning Deep Neural Networks under Label Noise in Remote Sensing Images
Neural Codes for Image Retrieval
A thorough discussion of several state-of-the-art techniques in image retrieval by considering the associated subproblems: image description, descriptor compression, nearest-neighbor search and query expansion, and the combined use of deep architectures and hand-crafted image representations for accurate and efficient image retrieval.
ImageNet Large Scale Visual Recognition Challenge
The creation of this benchmark dataset and the advances in object recognition that have been possible as a result are described, and the state-of-the-art computer vision accuracy with human accuracy is compared.
NoiseRank: Unsupervised Label Noise Reduction with Dependence Models
NoiseRank is proposed, for unsupervised label noise reduction using Markov Random Fields (MRF), which constructs a dependence model to estimate the posterior probability of an instance being incorrectly labeled given the dataset, and rank instances based on their estimated probabilities.