计算机工程与应用 ›› 2017, Vol. 53 ›› Issue (13): 34-41.DOI: 10.3778/j.issn.1002-8331.1703-0362

• 热点与综述 • 上一篇    下一篇

卷积神经网络在图像分类和目标检测应用综述

周俊宇,赵艳明   

  1. 中国传媒大学 理工学部,北京 100024
  • 出版日期:2017-07-01 发布日期:2017-07-12

Application of convolution neural network in image classification and object detection

ZHOU Junyu, ZHAO Yanming   

  1. College of Science and Technology, Communication University of China, Beijing 100024, China
  • Online:2017-07-01 Published:2017-07-12

摘要: 卷积神经网络具有强大的特征学习能力,随着大数据时代的到来和计算机能力的提升,近年来卷积神经网络在图像识别、目标检测等领域取得了突破性进展,掀起了新的研究热潮。综述卷积神经网络的基本原理,以及其在图像分类、目标检测上的研究进展和典型模型,最后分析了卷积神经网络目前的问题,并展望了未来的发展方向。

关键词: 卷积神经网络, 图像分类, 目标检测

Abstract: Convolutional Neural Network(CNN) has strong ability in finding characteristics of pictures. Recent years, with the arrival of big data era and the development of the computers, CNN has made great breakthrough in the field of image classification, object detection and so on. It has been widely studied in computer vision. This paper summarizes the theory of CNN, as well as the research and typical models on image classification, object detection. Finally, some existing problems in the CNN are analyzed, with forecasting the future development of CNN as well.

Key words: Convolutional Neural Network(CNN), image classification, object detection