计算机应用 ›› 2021, Vol. 41 ›› Issue (9): 2780-2784.DOI: 10.11772/j.issn.1001-9081.2020101533

所属专题: 前沿与综合应用

• 前沿与综合应用 • 上一篇    

基于改进DenseNet的牛眼图像特征提取方法

郑志强, 胡鑫, 翁智, 王雨禾, 程曦   

  1. 内蒙古大学 电子信息工程学院, 呼和浩特 010021
  • 收稿日期:2020-10-05 修回日期:2021-01-25 出版日期:2021-09-10 发布日期:2021-09-15
  • 通讯作者: 翁智
  • 作者简介:郑志强(1982-),男,内蒙古呼和浩特人,副教授,博士,主要研究方向:故障检测、机器学习、图像处理;胡鑫(1996-),女,内蒙古鄂尔多斯人,硕士,主要研究方向:图像处理;翁智(1978-),男,内蒙古乌兰察布人,教授,博士研究生,主要研究方向:人工智能、机器学习、图像处理;王雨禾(1992-),男(蒙古族),内蒙古呼和浩特人,硕士研究生,主要研究方向:机器视觉;程曦(1995-),女,内蒙古乌兰察布人,硕士研究生,主要研究方向:图像处理。
  • 基金资助:
    国家自然科学基金资助项目(61966026); 内蒙古自然科学基金资助项目(2020MS06015)。

Cattle eye image feature extraction method based on improved DenseNet

ZHENG Zhiqiang, HU Xin, WENG Zhi, WANG Yuhe, CHENG Xi   

  1. College of Electronic Information Engineering, Inner Mongolia University, Hohhot Inner Mongolia 010021, China
  • Received:2020-10-05 Revised:2021-01-25 Online:2021-09-10 Published:2021-09-15
  • Supported by:
    This work is partially supported by the National Natural Science Foundation of China (61966026), the Inner Mongolia Natural Science Foundation (2020MS06015).

摘要: 针对牛眼图像特征提取过程中由于梯度消失和过拟合造成的识别准确率较低的问题,提出一种基于改进DenseNet的牛眼图像特征提取方法。首先采用缩放指数线性单元(SeLU)激活函数防止网络梯度消失;其次通过DropBlock随机丢弃牛眼图像的特征块,从而防止过拟合,并加强网络的泛化能力;最后将改进后的稠密层进行叠加以组成改进的DenseNet(Dense convolutional Network)。在自建牛眼图像数据集上进行特征信息提取识别实验的结果表明,改进后的DenseNet识别准确率、精确率和召回率分别为97.47%、98.11%和97.90%;较改进前的网络在识别准确率、精确率和召回率上分别提升了2.52个百分点、3.32个百分点和2.94个百分点,可见改进后的网络具有较高的精度与鲁棒性。

关键词: 牛眼图像, 特征提取, 深度学习, DenseNet, DropBlock

Abstract: To address the problem of low recognition accuracy caused by vanishing gradient and overfitting in the cattle eye image feature extraction process, an improved DenseNet based cattle eye image feature extraction method was proposed. Firstly, the Scaled exponential Linear Unit (SeLU) activation function was used to prevent the vanishing gradient of the network. Secondly, the feature blocks of cattle eye images were randomly discarded by DropBlock, so as to prevent overfitting and strengthen the generalization ability of the network. Finally, the improved dense layers were superimposed to form an improved Dense convolutional Network (DenseNet). Feature information extraction recognition experiments were conducted on the self-built cattle eyes image dataset. Experimental results show that the recognition accuracy, precision and recall of the improved DenseNet are 97.47%, 98.11% and 97.90% respectively, and compared to the network without improvement, the above recognition accuracy rate, precision rate, recall rate are improved by 2.52 percentage points, 3.32 percentage points, 2.94 percentage points respectively. It can be seen that the improved network has higher precision and robustness.

Key words: cattle eye image, feature extraction, deep learning, DenseNet, DropBlock

中图分类号: