|Table of Contents|

[1] Wei Tao, Shuai Liguo, Zhang Yulu,. Influence of image data set noise on classificationwith a convolutional network [J]. Journal of Southeast University (English Edition), 2019, 35 (1): 51-56. [doi:10.3969/j.issn.1003-7985.2019.01.008]
Copy

Influence of image data set noise on classificationwith a convolutional network()
图像数据集噪声对卷积网络分类的影响
Share:

Journal of Southeast University (English Edition)[ISSN:1003-7985/CN:32-1325/N]

Volumn:
35
Issue:
2019 1
Page:
51-56
Research Field:
Computer Science and Engineering
Publishing date:
2019-03-30

Info

Title:
Influence of image data set noise on classificationwith a convolutional network
图像数据集噪声对卷积网络分类的影响
Author(s):
Wei Tao Shuai Liguo Zhang Yulu
School of Mechanical Engineering, Southeast University, Nanjing 211189, China
韦韬 帅立国 张雨露
东南大学机械工程学院, 南京 211189
Keywords:
image recognition data set noise deep convolutional network filtering of cross-category noise
图像识别 数据集噪声 深度卷积网络 跨类噪声筛选
PACS:
TP391.4
DOI:
10.3969/j.issn.1003-7985.2019.01.008
Abstract:
To evaluate the influence of data set noise, the network in network(NIN)model is introduced and the negative effects of different types and proportions of noise on deep convolutional models are studied. Different types and proportions of data noise are added to two reference data sets, Cifar-10 and Cifar-100. Then, this data containing noise is used to train deep convolutional models and classify the validation data set. The experimental results show that the noise in the data set has obvious adverse effects on deep convolutional network classification models. The adverse effects of random noise are small, but the cross-category noise among categories can significantly reduce the recognition ability of the model. Therefore, a solution is proposed to improve the quality of the data sets that are mixed into a single noise category. The model trained with a data set containing noise is used to evaluate the current training data and reclassify the categories of the anomalies to form a new data set. Repeating the above steps can greatly reduce the noise ratio, so the influence of cross-category noise can be effectively avoided.
为了评估数据集噪声的影响, 引入了NIN(network in network)模型, 研究了不同类型和比例的噪声对深度卷积模型的负面影响.将不同种类和比例的数据噪声加入基准数据集Cifar-10和Cifar-100, 然后使用这些包含噪声的数据来训练深度卷积模型, 并对验证数据集进行分类.实验结果表明, 数据集中的噪声对深度卷积网络分类模型确实有明显的不利影响.其中, 随机噪声的不利影响很小, 但是类别之间的跨类噪声却显著地降低了模型的识别能力.因此, 提出了一种解决方案用来改进混入单类别噪声的数据集质量, 即用含有噪声的数据集训练的模型评价当前训练数据, 并将异常的类别重新归类以形成新的数据集.经过多轮迭代训练, 可以大大降低其噪声比率, 从而可以有效避免交叉类别噪声的影响.

References:

[1] Erhan D, Courville A, Bengio Y. Understanding representations learned in deep architectures, Technical Report 1355 [R]. Montreal, Canada: Universite de Montr’ eal, 2010.
[2] Hinton G E. Reducing the dimensionality of data with neural networks[J].Science, 2006, 313(5786): 504-507. DOI:10.1126/science.1127647.
[3] Krizhevsky A, Hinton G. Learning multiple layers of features from tiny images[R]. Toronto, Ontario, Canada: University of Toronto, 2009.
[4] Yu D, Deng L. Deep learning and its applications to signal and information processing [exploratory DSP][J]. IEEE Signal Processing Magazine, 2011, 28(1): 145-154. DOI:10.1109/msp.2010.939038.
[5] Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks[C]// International Conference on Neural Information Processing Systems. Curran Associates Inc, 2012:1-9.
[6] Springenberg J T, Dosovitskiy A, Brox T, et al. Striving for simplicity: The all convolutional net[J]. arXiv preprint arXiv:1412.6806, 2015.
[7] Zagoruyko S, Komodakis N. Wide residual networks[J]. arXiv preprint arXiv:1605.07146, 2016.
[8] Szegedy C, Vanhoucke V, Ioffe S, et al. Rethinking the inception architecture for computer vision[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2016: 2818-2826.
[9] Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition[J]. arXiv preprint arXiv:1409.1556, 2014.
[10] Lin M, Chen Q, Yan S. Network in network[J]. arXiv preprint arXiv:1312.4400, 2013.

Memo

Memo:
Biographies: Wei Tao(1995—), male, graduate; Shuai Liguo(corresponding author), male, doctor, professor, liguo.shuai@126.com.
Foundation item: The Science and Technology R&D Fund Project of Shenzhen(No.JCYJ2017081765149850).
Citation: Wei Tao, Shuai Liguo, Zhang Yulu.Influence of image data set noise on classification with a convolutional network[J].Journal of Southeast University(English Edition), 2019, 35(1):51-56.DOI:10.3969/j.issn.1003-7985.2019.01.008.
Last Update: 2019-03-20