Abstract:Summary The performance of supervised deep learning-based denoising methods relies on massive amounts of high-quality training dataset with labels. However, data labeling is a time-consuming and tedious process, and the lack of labeled dataset has become a major bottleneck affecting the development of supervised deep learning-based denoising methods. In recent years, denoising methods that only use unlabeled noisy dataset for training have received more and more attention. Although these methods get rid of the dependence on labels, they usually have some specific requirements on the training dataset. For example, the paired training data are required to be multiple noisy observations for each scene or obey a specific noise distribution, etc., which are often very challenging to meet in practical applications. In this study, we propose an improved self-supervised denoising framework based on noise2noise that only uses noisy seismic dataset for training, and we name it sample2sample. The proposed denoising framework does not require multiple repeated acquisitions of seismic data to obtain multiple independent noisy observations for each scene used for training, and has no specific requirement for the noise distribution prior. Specifically, we introduce a random sampler to generate paired subsamples from some individual noisy seismic data for training. The corresponding elements in the two paired subsamples are adjacent in the original seismic data and approximately meet the training premise of noise2noise, i.e. the paired training data have the same signal. In addition, considering that there are some subtle differences in the signals of the paired subsamples generated by sampling, we also introduce a regularization loss to compensate for this. We conducted a qualitative and quantitative analysis of the denoising performance of the proposed method through further experiments, including synthetic data experiments and field data experiments.

Sample Reweighting for Label Denoising of Neural Activity Data

Meta-Learning for Decoding Neural Activity Data with Noisy Labels

Noise is the Fatal Poison: A Noise-aware Network for Noisy Dataset Classification

Making Deep Neural Networks Robust to Label Noise: A Reweighting Loss and Data Filtration

Learning from Noisy Labels with Decoupled Meta Label Purifier

Learning with Noisy Labels Via Self-supervised Adversarial Noisy Masking

Learning from Noisy Labels with Coarse-to-Fine Sample Credibility Modeling

VRLVMix: Combating Noisy Labels with Sample Selection Based on Loss Variation

Denoising Noisy Neural Networks: A Bayesian Approach with Compensation.

Meta-probability Weighting for Improving Reliability of DNNs to Label Noise

Sample2Sample: an Improved Self-Supervised Denoising Framework for Random Noise suppression in distributed acoustic sensing vertical seismic profile data

Noisy Label Learning Based on Weighted Neighborhood Consistency

INN: A Method Identifying Clean-annotated Samples via Consistency Effect in Deep Neural Networks

Regroup Median Loss for Combating Label Noise

MetaInfoNet: Learning Task-Guided Information for Sample Reweighting

Learning to Reweight Samples with Offline Loss Sequence

Learning from Noisy Labels via Self-Taught On-the-Fly Meta Loss Rescaling

Self-cognitive Denoising in the Presence of Multiple Noisy Label Sources

On Better Detecting and Leveraging Noisy Samples for Learning with Severe Label Noise

Bayesian statistics guided label refurbishment mechanism: Mitigating label noise in medical image classification

Neighborhood Collective Estimation for Noisy Label Identification and Correction