面向网络内容安全的图像识别技术研究  

Research on Image Recognition Technology for the Network Content Security

作　　者：崔鹏飞 裘玥[2] 孙瑞[2] 

机构地区：[1]郑州金惠计算机系统工程有限公司,河南郑州450001 [2]北京市公安局网络安全保卫总队,北京100740

出　　处：《信息网络安全》2015年第9期154-157,共4页Netinfo Security

摘　　要：随着网络技术的快速发展,大量的违法不良图像在网络中大肆传播,严重危害了网络内容安全。文章从网络安全工作的实际出发,针对传统文本过滤技术的缺点,指出了图像识别技术对于过滤违法不良信息的重要性。文章分析总结了几种违法不良图像识别技术,并探讨了每种技术的适用场景。最后,结合图像识别的本质问题,指出深度学习是未来图像识别技术的发展趋势,讨论了利用卷积神经网络代替传统的方法对违法不良图像进行识别的方法。With the rapid development of network technology, a variety of illegal images spread suddenly in the network, which endangers the network content security seriously. From the reality of network security in this paper, aiming at the shortcomings of the traditional text ifltering technology, this paper point out the importance of the image recognition technology for ifltering the illegal and bad information. The paper analyzes and summarizes some kinds of recognition technologies of the illegal images, and discusses the application scene of each technology. Finally, combining with the essence of image recognition problems, the paper indicate that deep learning is the development trend of the image recognition technology in the future, and discuss how to use convolution neural network instead of the traditional method to recognize the illegal images.

关 键 词：图像识别 网络内容安全 深度学习 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术] TP391.41[自动化与计算机技术—计算机科学与技术]