Noise-Suppression Deep Learning Strategy via Reliability Estimation Gate for Image Classification

SHAO Hang; HUANG Hai-liang; GUO Yu-chen; DAI Qiong-hai

doi:10.3969/j.issn.0372-2112.2020.10.014

您当前的位置：

首页 >

文章列表页 >

Noise-Suppression Deep Learning Strategy via Reliability Estimation Gate for Image Classification

更新时间：2023-02-24

- Noise-Suppression Deep Learning Strategy via Reliability Estimation Gate for Image Classification
- Acta Electronica Sinica Vol. 48, Issue 10, Pages: 1969-1975(2020)
- 作者机构：
  
  1. 浙江未来技术研究院(嘉兴),浙江,嘉兴,314000
  3. 清华大学自动化系,北京,100084
- 作者简介：
- 基金信息：
- DOI：10.3969/j.issn.0372-2112.2020.10.014
  CLC： TP301.6
- Published Online：25 October 2020，
  
  Published：2020
- 稿件说明：
移动端阅览
SHAO Hang, HUANG Hai-liang, GUO Yu-chen, et al. Noise-Suppression Deep Learning Strategy via Reliability Estimation Gate for Image Classification[J]. Acta Electronica Sinica, 2020, 48(10): 1969-1975.
DOI：

SHAO Hang, HUANG Hai-liang, GUO Yu-chen, et al. Noise-Suppression Deep Learning Strategy via Reliability Estimation Gate for Image Classification[J]. Acta Electronica Sinica, 2020, 48(10): 1969-1975. DOI： 10.3969/j.issn.0372-2112.2020.10.014.

摘要

近年来，深度学习越来越关注噪声抑制的研究.本文提出了一种噪声抑制深度学习策略，该策略通过构建噪声无感网络（Noise Unaware Network，NUN）和可信度估计单元（Reliability Estimation Gate，REG）来处理训练数据含有噪声的情况.通过对每个样本的可信度进行评估，调节其在训练时的权重，从而降低标签噪声对网络训练的影响.随着模型的迭代更新，标签可信数据的权重将会逐渐变大，而噪声数据的权重将会被抑制.本文通过在多个标注数据集上的实验验证了所提出的噪声抑制深度学习策略的有效性.

Abstract

In recent years

more and more attention has been paid to the study and research of noise suppression.This paper proposes a noise suppression deep learning strategy that deals with the presence of noise in training data by constructing a Noise Unaware Network (NUN) and Reliability Estimation Gate (REG).By evaluating the reliability of each sample and adjusting its weight during training

the influence of label noise on network training can be reduced.As the model is updated iteratively

the weight of clean data will gradually increase

while the weight of noise data will be suppressed.Experiments on multiple benchmark data sets demonstrate the effectiveness of the proposed deep learning strategy for noise suppression.

关键词

Keywords

references

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Research of Lightweight Convolution Neural Network Based on Feature Expansion Convolution

Variable Horizon Multi-Directional Scanning Method for Time Series Anomaly Detection

Differentially Private with Sparse and Smooth Self-Distillation

Operator Fusion Method and Hardware Architecture Design Based on Non-Standard Operators

Related Author

XU Xin-zheng

LI Shan

Xin-zheng XU

Shan LI

HUANG Yu-zhe

GUAN Yong-yuan

WEI Song-jie

ZHAO Deng-feng

Related Institution

School of Computer Science and Technology, China University of Mining and Technology

Engineering Research Center of Mining Digital Ministry of Education

School of Computer Science and Technology， China University of Mining and Technology

School of Computer Science and Engineering, School of Cyber Science and Engineering, Nanjing University of Science and Technology

School of Information, Renmin University of China

⁰