摘要
卷积神经网络是人工神经网络与深度学习相结合,从而实现深度学习的方法.其具有良好的容错性、自适应性以及较强的自学习能力,还具有自动提取特征、权值共享以及输入图像与网络结构结合良好等优势.基于卷积神经网络和深度学习的图像识别系统,首先对不同的图像进行采集,将采集的得到的结果作为训练集和测试集.通过卷积神经网络对采集结果的训练,得到用来识别的各类特征,识别的结果可以得到图像的类别信息.
Convolutional neural networks are a combination of artificial neural networks and deep learning to achieve deep learning.It has good fault tolerance,adaptability and strong self-learning ability.It also has the advantages of automatic feature extraction,weight sharing and good combination of input image and network structure.Based on the convolutional neural network and deep learning image recognition system,different images are first collected,and the obtained results are used as a training set and a test set.Through the training of the collected results by the convolutional neural network,various types of features for identification are obtained,and the re sult of the recognition can obtain the category information of the image.
作者
李航
厉丹
朱晨
姚瑶
张丽娜
LI Hang;LI Dan;ZHU Chen;YAO Yao;ZHANG Li-na(Information and Electrical Engineering College,Xuzhou Institute of Technology,Xuzhou 221000,China)
出处
《电脑知识与技术》
2020年第10期196-197,200,共3页
Computer Knowledge and Technology
基金
徐州市科技计划项目(KC18011)
徐州工程学院大学生创新创业训练计划项目(xcx2019031)
江苏省教育信息化研究课题(20180071)
徐州工程学院高等教育研究课题(YGJ1916)
江苏省现代教育课题(2019-R-69623)
2018年第二批产学合作协同育人项目。
关键词
卷积层神经网络
深度学习
图像识别
图像分类
Alexnet构架
convolutional neural network
deep learning
image identification
image classification
Alexnet architecture