• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

计算机工程与科学

• 图形与图像 • 上一篇    下一篇

基于卷积神经网络的图像数据增强算法

蒋芸,张海,陈莉,陶生鑫   

  1. (西北师范大学计算机科学与工程学院,甘肃 兰州 730070)
  • 收稿日期:2019-04-29 修回日期:2019-06-04 出版日期:2019-11-25 发布日期:2019-11-25
  • 基金资助:

    国家自然科学基金(61962054);2016年甘肃省科技计划资助自然科学基金(1606RJZA047);2012年度甘肃省高校基本科研业务费专项资金;甘肃省高校研究生导师项目(1201-16);西北师范大学第三期知识与创新工程科研骨干项目(nwnu-kjcxgc-03-67)

An image data augmentation algorithm
based on convolutional neural networks

JIANG Yun,ZHANG Hai,CHEN Li,TAO Sheng-xin   

  1. (College of Computer Science and Engineering,Northwest Normal University,Lanzhou 730070,China)
  • Received:2019-04-29 Revised:2019-06-04 Online:2019-11-25 Published:2019-11-25

摘要:

提升卷积神经网络的泛化能力和降低过拟合的风险是深度卷积神经网络的研究重点。遮挡是影响卷积神经网络泛化能力的关键因素之一,通常希望经过复杂训练得到的模型能够对遮挡图像有良好的泛化性。为了降低模型过拟合的风险和提升模型对随机遮挡图像识别的鲁棒性,提出了激活区域处理算法,
在训练过程中对某一卷积层的最大激活特征图进行处理后对输入图像进行遮挡,然后将被遮挡的新图像作为网络的新输入并继续训练模型。实验结果表明,提出的算法能够提高多种卷积神经网络模型在不同数据集上的分类性能,并且训练好的模型对随机遮挡图像的识别具有非常好的鲁棒性。
 
 

关键词: 深度学习, 卷积神经网络, 图像分类, 数据增强

Abstract:

Improving the generalization ability and reducing the over-fitting risk is the research focus of deep convolutional neural networks. Occlusion is one of the critical factors affecting the generalization ability of convolutional neural networks. It is usually hoped that the models after complex training can have a good generalization for occlusion images.In order to reduce the over-fitting risk and improve the robustness of the model to random occlusion image recognition, this paper proposes an activation feature processing algorithm. During the training process, the input image is occluded by processing the maximum activation feature map of a convolutional layer, then the occluded new image is used as a new input to the network to go on training the model. The experimental results show that the proposed algorithm can improve the classification performance of multiple convolutional neural network models on different datasets and the trained models have excellent robustness to the identification of random occlusion images.
 

Key words: deep learning, convolutional neural network, image classification, data augmentation