Noisy Label Processing for Classification: A Survey

TL;DR

This survey reviews deep learning methods for handling noisy labels in image classification, discusses noise patterns, and introduces a new synthetic benchmark based on real-world data to evaluate noise-robust algorithms.

Contribution

It provides a comprehensive review of approaches for noisy label mitigation and proposes a novel data-guided synthetic noise generation method and benchmark.

Findings

Analysis of various noise-robust methods' performance

Introduction of a real-world data-guided synthetic noise benchmark

Insights into noise patterns in real-world datasets

Abstract

In recent years, deep neural networks (DNNs) have gained remarkable achievement in computer vision tasks, and the success of DNNs often depends greatly on the richness of data. However, the acquisition process of data and high-quality ground truth requires a lot of manpower and money. In the long, tedious process of data annotation, annotators are prone to make mistakes, resulting in incorrect labels of images, i.e., noisy labels. The emergence of noisy labels is inevitable. Moreover, since research shows that DNNs can easily fit noisy labels, the existence of noisy labels will cause significant damage to the model training process. Therefore, it is crucial to combat noisy labels for computer vision tasks, especially for classification tasks. In this survey, we first comprehensively review the evolution of different deep learning approaches for noisy label combating in the image classification task. In addition, we also review different noise patterns that have been proposed to design robust algorithms. Furthermore, we explore the inner pattern of real-world label noise and propose an algorithm to generate a synthetic label noise pattern guided by real-world data. We test the algorithm on the well-known real-world dataset CIFAR-10N to form a new real-world data-guided synthetic benchmark and evaluate some typical noise-robust methods on the benchmark.