基于改进神经网络的图像识别方法研究  

Research on Image Recognition Method Based on Improved Neural Network

作　　者：王一竹[1] WANG Yizhu(Wuxi Institute of Technology,Wuxi Jiangsu 214121,China)

机构地区：[1]无锡职业技术学院,江苏无锡214121

出　　处：《信息与电脑》2021年第18期185-187,共3页Information & Computer

摘　　要：传统的机器学习图像识别方法存在很多缺陷,识别图像的正确率低,因此基于改进神经网络设计新的图像识别方法,首先进行图像预处理,然后选择了改进神经网络参数,最后构建了图像识别模型,从而实现图像识别。实验结果表明:设计的图像识别方法的准确率更高,具有一定的应用价值。Traditional machine learning image recognition methods have many shortcomings, and the accuracy of recognizing images is low. Therefore, a new image recognition method is designed based on improved neural networks. First, image preprocessing is performed, then improved neural network parameters are selected, and finally image recognition is constructed. The experimental results show that the designed image recognition method is more accurate and has certain application value.

关 键 词：改进神经网络 图像识别 深度学习 图像预处理 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]