CleanNet: Transfer Learning for Scalable Image Classifier Training with Label Noise

Abstract

In this paper, we study the problem of learning image classification models with label noise. Existing approaches depending on human supervision are generally not scalable as manually identifying correct or incorrect labels is time-consuming, whereas approaches not relying on human supervision are scalable but less effective. To reduce the amount of human supervision for label noise cleaning, we introduce CleanNet, a joint neural embedding network, which only requires a fraction of the classes being manually verified to provide the knowledge of label noise that can be transferred to other classes. We further integrate CleanNet and conventional convolutional neural network classifier into one framework for image classification learning. We demonstrate the effectiveness of the proposed algorithm on both of the label noise detection task and the image classification on noisy data task on several large-scale datasets. Experimental results show that CleanNet can reduce label noise detection error rate on held-out classes where no human supervision available by 41.5% compared to current weakly supervised methods. It also achieves 47% of the performance gain of verifying all images with only 3.2% images verified on an image classification task. Source code and dataset will be available at kuanghuei.github.io/CleanNetProject.

arXiv commentAccepted to CVPR 2018

References in corpus (10)

• Scikit-learn: Machine Learning in Python
• LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loop
• Matching Networks for One Shot Learning
• Training Convolutional Networks with Noisy Labels
• Adversarial Discriminative Domain Adaptation
• WebVision Database: Visual Learning and Understanding from Web Data
• Webly Supervised Learning of Convolutional Networks
• Learning from Noisy Labels with Distillation
• Learning From Noisy Large-Scale Datasets With Minimal Supervision
• Learning Robust Visual-Semantic Embeddings

Cited by in corpus (7)

• Symmetric Cross Entropy for Robust Learning with Noisy Labels
• Stacked Cross Attention for Image-Text Matching
• Image Captioning with Very Scarce Supervised Data: Adversarial Semi-Supervised Learning Approach
• Weakly Supervised Learning with Side Information for Noisy Labeled Images
• Joint Negative and Positive Learning for Noisy Labels
• NLNL: Negative Learning for Noisy Labels
• Product Image Recognition with Guidance Learning and Noisy Supervision