摘要
针对传统的图像分类算法忽略图像多个对象之间的关系,同时存在人类感知高层语义信息和底层图像特征表达之间的障碍等不足,引入了基于深度卷积神经网络的人脸图像识别算法.该算法借助于深度学习,对经典的卷积神经网络分别从内部结构和网络框架上进行优化和改进,通过增加网络结构深度和优化训练模型提取出图像高层语义特征,继而提高图像分类的精确度.实验表明,改进后的深度卷积神经网络分类算法具有良好的有效性和鲁棒性.
The traditional image classification algorithm ignores the relationship between multiple objects in the image,at the same time,there is an obstacle between human perception of high-level semantic information and underlying image feature expression.An image recognition algorithm based on deep convolutional neural network was proposed,the algorithm uses deep learning to optimize and improve the classic convolutional neural network from the internal structure and network framework respectively,by increasing the network structure depth and optimizing the training model,the high-level semantic features of the image were extracted,which in turn improves the accuracy of image classification.Experiments showed that image classification algorithm based on deep convolutional neural network had better classification effect.
作者
陈瑞瑞
CHEN Ruirui(College of Information Engineering,Zhengzhou University of Industrial Technology,Zhengzhou 451100,China)
出处
《河南科技学院学报（自然科学版）》
2018年第4期50-54,61,共6页
Journal of Henan Institute of Science and Technology(Natural Science Edition)
基金
河南省科技攻关计划项目(162102210119)
关键词
深度学习
卷积神经网络
图像分类
deep learning
convolutional neural network
image classification
