Image Classification And Recognition Based On The Deep Convolutional Neural Network

摘要：

　　With the development of the information age,there were a lot of data whose features couldnt be extracted or predicted effectively in real life.One of the core function of computer vision technology is to classify and recognize,with classification and recognition as its summit mission of object detection and object positioning.Due to image data were affected by multiple factors such as illumination,environment,angle,certain object features couldnt be established by manual coding,and it is hard for high latitude data in a computer to realize real-time object detection,object localization,classification and recognition.Therefore the higher accuracy in classification could be obtained with the help of GPU of high performance and the large scale pretraining on the super database Image Net,based on the most advanced deep convolution neural network algorithm.The complete optimization training was conducted on data sets Pascal Vision Object Classes(VOC),and the real-time object detection,object localization,classification and recognition were realized by high performance GPU of NVIDIA.

关键词： Machine Learning Computer Vision Deep Learning Convolutional Neural Network

作者: Yuan-yuan WANG Long-jun ZHANG Yang XIAO Jing XU You-jun ZHANG

作者单位: The College of Information and Electric Engineering,Shenyang Agricultural University,Shenyang,110866 STATE GRID XINJIANG INFORMATION & TELECOMMUNICATION COMPANY,830000,China

会议类型: 国际会议

会议名称: 2017年第2届联合国际信息技术、机械与电子工程国际会议(JIMEC2017)

会议地点: 重庆

会议语种:英文

页码: 171-174

在线出版日期: 2017-10-04（万方平台首次上网日期，不代表论文的发表时间）