Image Classification with Deep Learning in the Presence of Noisy Labels: A Survey

Abstract

Image classification systems recently made a giant leap with the advancement of deep neural networks. However, these systems require an excessive amount of labeled data to be adequately trained. Gathering a correctly annotated dataset is not always feasible due to several factors, such as the expensiveness of the labeling process or difficulty of correctly classifying data, even for the experts. Because of these practical challenges, label noise is a common problem in real-world datasets, and numerous methods to train deep neural networks with label noise are proposed in the literature. Although deep neural networks are known to be relatively robust to label noise, their tendency to overfit data makes them vulnerable to memorizing even random noise. Therefore, it is crucial to consider the existence of label noise and develop counter algorithms to fade away its adverse effects to train deep neural networks efficiently. Even though an extensive survey of machine learning techniques under label noise exists, the literature lacks a comprehensive survey of methodologies centered explicitly around deep learning in the presence of noisy labels. This paper aims to present these algorithms while categorizing them into one of the two subgroups: noise model based and noise model free methods. Algorithms in the first group aim to estimate the noise structure and use this information to avoid the adverse effects of noisy labels. Differently, methods in the second group try to come up with inherently noise robust algorithms by using approaches like robust losses, regularizers or other learning paradigms.

References in corpus (44)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• Distilling the Knowledge in a Neural Network
• Explaining and Harnessing Adversarial Examples
• Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• mixup: Beyond Empirical Risk Minimization
• Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results
• LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loop
• Learning to Reweight Examples for Robust Deep Learning
• Classification with Noisy Labels by Importance Reweighting
• Semi-Supervised Learning with Ladder Networks
• Training Deep Neural Networks on Noisy Labels with Bootstrapping
• DivideMix: Learning with Noisy Labels as Semi-supervised Learning
• Using Pre-Training Can Improve Model Robustness and Uncertainty
• Deep Learning is Robust to Massive Label Noise
• Regularizing Neural Networks by Penalizing Confident Output Distributions
• A Closer Look at Memorization in Deep Networks
• WebVision Database: Visual Learning and Understanding from Web Data
• Noise Tolerance under Risk Minimization
• Dimensionality-Driven Learning with Noisy Labels
• Making Risk Minimization Tolerant to Label Noise
• Hyperspectral Image Classification in the Presence of Noisy Labels
• Are Anchor Points Really Indispensable in Label-Noise Learning?
• Masking: A New Perspective of Noisy Supervision
• Data Poisoning Attacks on Factorization-Based Collaborative Filtering
• Can Image-Level Labels Replace Pixel-Level Labels for Image Parsing
• Curriculum Loss: Robust Learning and Generalization against Label Corruption
• Learning From Noisy Singly-labeled Data
• Combating Label Noise in Deep Learning Using Abstention
• Understanding and Utilizing Deep Neural Networks Trained with Noisy Labels
• Robustness of Conditional GANs to Noisy Labels
• On Symmetric Losses for Learning from Corrupted Labels
• Loss factorization, weakly supervised learning and label noise robustness
• Label Noise Types and Their Effects on Deep Learning
• Error-Bounded Correction of Noisy Labels
• Learning to Learn from Weak Supervision by Full Supervision
• Avoiding Your Teacher's Mistakes: Training Neural Networks with Controlled Weak Supervision
• Boosting in the presence of label noise
• Robust Learning Under Label Noise With Iterative Noise-Filtering
• Improving Multi-Person Pose Estimation using Label Correction
• SOSELETO: A Unified Approach to Transfer Learning and Training with Noisy Labels
• Safeguarded Dynamic Label Regression for Generalized Noisy Supervision
• Photometric Transformer Networks and Label Adjustment for Breast Density Prediction

Cited by in corpus (35)

• A Survey of Label-noise Representation Learning: Past, Present and Future
• Toward Open-World Electroencephalogram Decoding Via Deep Learning: A Comprehensive Survey
• Embrace Limited and Imperfect Training Datasets: Opportunities and Challenges in Plant Disease Recognition Using Deep Learning
• Identifying Mislabeled Data using the Area Under the Margin Ranking
• Unsupervised classification to improve the quality of a bird song recording dataset
• Label Noise Types and Their Effects on Deep Learning
• How quantum computing can enhance biomarker discovery
• How Accurate Does It Feel? -- Human Perception of Different Types of Classification Mistakes
• MetaLabelNet: Learning to Generate Soft-Labels from Noisy-Labels
• One-Step Abductive Multi-Target Learning with Diverse Noisy Samples and Its Application to Tumour Segmentation for Breast Cancer
• Beyond Cats and Dogs: Semi-supervised Classification of fuzzy labels with overclustering
• Fuzzy Overclustering: Semi-Supervised Classification of Fuzzy Labels with Overclustering and Inverse Cross-Entropy
• Dive into Ambiguity: Latent Distribution Mining and Pairwise Uncertainty Estimation for Facial Expression Recognition
• Towards Understanding Deep Learning from Noisy Labels with Small-Loss Criterion
• Webly Supervised Image Classification with Metadata: Automatic Noisy Label Correction via Visual-Semantic Graph
• A Study on the Impact of Data Augmentation for Training Convolutional Neural Networks in the Presence of Noisy Labels
• Generative models-based data labeling for deep networks regression: application to seed maturity estimation from UAV multispectral images
• EfficientCLIP: Efficient Cross-Modal Pre-training by Ensemble Confident Learning and Language Modeling
• Deep Learning from Small Amount of Medical Data with Noisy Labels: A Meta-Learning Approach
• Impact of imperfect annotations on CNN training and performance for instance segmentation and classification in digital pathology
• Distant Supervision and Noisy Label Learning for Low Resource Named Entity Recognition: A Study on Hausa and Yorùbá
• Labels Are Not Perfect: Inferring Spatial Uncertainty in Object Detection
• Negative-ResNet: Noisy Ambulatory Electrocardiogram Signal Classification Scheme
• Empirical Perspectives on One-Shot Semi-supervised Learning
• Learning from Ambiguous Labels for Lung Nodule Malignancy Prediction
• Multi-Label Gold Asymmetric Loss Correction with Single-Label Regulators
• Rethinking Noisy Label Models: Labeler-Dependent Noise with Adversarial Awareness
• Self-Supervised Person Detection in 2D Range Data using a Calibrated Camera
• Learning advisor networks for noisy image classification
• An Exploration into why Output Regularization Mitigates Label Noise
• Semi-Supervised Multi-Task Learning for Interpretable Quality As- sessment of Fundus Images
• FISHing in Uncertainty: Synthetic Contrastive Learning for Genetic Aberration Detection
• Analysis of classifiers robust to noisy labels
• Do We Really Need Gold Samples for Sample Weighting Under Label Noise?
• A Symmetric Loss Perspective of Reliable Machine Learning