Deep Learning From Noisy Image Labels With Quality Embedding

Yao, Jiangchao; Wang, Jiajie; Tsang, Ivor W.; Zhang, Ya; Sun, Jun; Zhang, Chengqi; Zhang, Rui

doi:10.1109/tip.2018.2877939

Cited by 101 publications

(45 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Both models are, in general, comparable to competing architectures in terms of overall performance. These results provide evidence that deep learning algorithms trained on a sufficiently large dataset are robust against a moderate level of label noise, which corresponds with findings from previously reported studies (Northcutt et al, 2017;Rolnick et al, 2017;Yao et al, 2017).…”

Section: Lesion Classification Performancesupporting

confidence: 91%

Learning to detect chest radiographs containing pulmonary lesions using visual attention networks

Pesce

Withey

Ypsilantis

et al. 2019

Medical Image Analysis

106

View full text Add to dashboard Cite

Machine learning approaches hold great potential for the automated detection of lung nodules on chest radiographs, but training algorithms requires very large amounts of manually annotated radiographs, which are difficult to obtain. The increasing availability of PACS (Picture Archiving and Communication System), is laying the technological foundations needed to make available large volumes of clinical data and images from hospital archives. Binary labels indicating whether a radiograph contains a pulmonary lesion can be extracted at scale, using natural language processing algorithms. In this study, we propose two novel neural networks for the detection of chest radiographs containing pulmonary lesions. Both architectures make use of a large number of weakly-labelled images combined with a smaller number of manually annotated x-rays. The annotated lesions are used during training to deliver a type of visual attention feedback informing the networks about their lesion localisation performance. The first architecture extracts saliency maps from high-level convolutional layers and compares the inferred position of a lesion against the true position when this information is available; a localisation error is then back-propagated along with the softmax classification error. The second approach consists of a recurrent attention model that learns to observe a short sequence of smaller image portions through reinforcement learning; the reward function penalises the exploration of areas, within an image, that are unlikely to contain nodules. Using a repository of over 430,000 historical chest radiographs, we present and discuss the proposed methods over related architectures that use either weakly-labelled or annotated images only.

show abstract

Section: Lesion Classification Performancesupporting

confidence: 91%

Learning to detect chest radiographs containing pulmonary lesions using visual attention networks

Pesce

Withey

Ypsilantis

et al. 2019

Medical Image Analysis

106

View full text Add to dashboard Cite

show abstract

“…However, existing methods often have prerequisites that may not be practical in many applications, e.g., an auxiliary set with clean labels [28] or prior information about the noise [16]. Some methods are very complex [29], which hurts their deployment capability. Overfitting to noise is another serious difficulty.…”

Section: Introductionmentioning

confidence: 99%

Probabilistic End-To-End Noise Correction for Learning With Noisy Labels

2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

351

249

View full text Add to dashboard Cite

Deep learning has achieved excellent performance in various computer vision tasks, but requires a lot of training examples with clean labels. It is easy to collect a dataset with noisy labels, but such noise makes networks overfit seriously and accuracies drop dramatically. To address this problem, we propose an end-to-end framework called PEN-CIL, which can update both network parameters and label estimations as label distributions. PENCIL is independent of the backbone network structure and does not need an auxiliary clean dataset or prior information about noise, thus it is more general and robust than existing methods and is easy to apply. PENCIL outperforms previous state-of-the-art methods by large margins on both synthetic and real-world datasets with different noise types and noise rates. Experiments show that PENCIL is robust on clean datasets, too.

show abstract

“…Subsequent improvement in (Goldberger and Ben-Reuven 2017) model the noise transition with a Softmax layer and tune its parameters along with the training progress. Based on this work, Yao et al(2017) introduced an auxiliary variable to augment the noise transition with more uncertainty. Han et al(2018a) further added the structure information to constrain the op-timization.…”

Section: Related Workmentioning

confidence: 99%

Safeguarded Dynamic Label Regression for Noisy Supervision

Yao

Zhang

et al. 2019

AAAI

Self Cite

View full text Add to dashboard Cite

Learning with noisy labels is imperative in the Big Data era since it reduces expensive labor on accurate annotations. Previous method, learning with noise transition, has enjoyed theoretical guarantees when it is applied to the scenario with the class-conditional noise. However, this approach critically depends on an accurate pre-estimated noise transition, which is usually impractical. Subsequent improvement adapts the preestimation in the form of a Softmax layer along with the training progress. However, the parameters in the Softmax layer are highly tweaked for the fragile performance and easily get stuck into undesired local minimums. To overcome this issue, we propose a Latent Class-Conditional Noise model (LCCN) that models the noise transition in a Bayesian form. By projecting the noise transition into a Dirichlet-distributed space, the learning is constrained on a simplex instead of some adhoc parametric space. Furthermore, we specially deduce a dynamic label regression method for LCCN to iteratively infer the latent true labels and jointly train the classifier and model the noise. Our approach theoretically safeguards the bounded update of the noise transition, which avoids arbitrarily tuning via a batch of samples. Extensive experiments have been conducted on controllable noise data with CIFAR10 and CIFAR-100 datasets, and the agnostic noise data with Clothing1M and WebVision17 datasets. Experimental results have demonstrated that the proposed model outperforms several state-of-the-art methods.

show abstract

Deep Learning From Noisy Image Labels With Quality Embedding

Cited by 101 publications

References 28 publications

Learning to detect chest radiographs containing pulmonary lesions using visual attention networks

Learning to detect chest radiographs containing pulmonary lesions using visual attention networks

Probabilistic End-To-End Noise Correction for Learning With Noisy Labels

Safeguarded Dynamic Label Regression for Noisy Supervision

Contact Info

Product

Resources

About