高维数据集在卷积神经网络中的应用研究  

Research on the application of convolutional neural network in high-dimensional datasets

作　　者：罗泽鹏 范峰 高宇航 Luo Zepeng;Fan Feng;Gao Yuhang(North China University of Science and Technology,Tangshan 063210,China)

机构地区：[1]华北理工大学,河北唐山063210

出　　处：《无线互联科技》2018年第13期95-98,共4页Wireless Internet Technology

摘　　要：大规模的模拟和传感器设备阵列与现今日益庞大的计算资源相结合,产生了海量的复杂高维数据集,这些数据集应用在众多领域的研究中。与此同时,计算机视觉与模式识别领域的研究者们取得了许多重大突破,尤其是卷积神经网络近年来在该领域取得的成就尤为突出。文章将高维数据应用于深度卷积神经网络中,通过实验与分析比对,证明高维数据应用于卷积神经网络中的可能性。The combination of large-scale arrays of analog and sensor devices and the ever-increasing number of computational resources available today has resulted in a vast array of complex high-dimensional datasets that are used in numerous fields of research. At the same time, researchers in the field of computer vision and pattern recognition have made many major breakthroughs. In particular, convolutional neural networks have made outstanding achievements in this field in recent years. This paper focus on applying high-dimensional data to deep convolutional neural networks. Through the comparison of experiments and analysis, it proves the possibility of applying highdimensional data to convolutional neural networks.

关 键 词：数据可视化 深度学习 卷积神经网络 计算机视觉 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP391.41[自动化与计算机技术—控制科学与工程]