现代电子技术2017，Vol.40Issue(4)：12-15,4.DOI:10.16652/j.issn.1004-373x.2017.04.004
基于改进卷积神经网络的人体检测研究
Research on pedestrian detection based on improved convolutional neural network
摘要
Abstract
The deep learning algorithm was introduced to execute the human body feature extraction and pedestrian detec⁃tion because of the low performance of pedestrian detection histogram of oriented gradient in the complex background. The con⁃tent⁃based image retrieval method is used for data expansion to reduce the quantity demand of the training samples of convolu⁃tional neural network. The method is able to ensure the original database background distribution and pedestrian resolution. The Fisher criterion is imported when the reflection propagation weights of the convolutional neural network are updated in order to improve the detection efficiency of the algorithm. The back propagation algorithm is adopted to obtain the weight values of the in⁃ter⁃class constraint function in sample to ensure the classification accuracy while the errors exist. The test results on the INRIA database show that the omission rate and the detection rate of the improved algorithm have been improved,and can detect pedes⁃trians in the most complex⁃backgrounds successfully.关键词
行人检测/深度学习/卷积神经网络/复杂背景Key words
pedestrian detection/deep learning/convolutional neural network/complex background分类
信息技术与安全科学引用本文复制引用
左艳丽,马志强,左宪禹..基于改进卷积神经网络的人体检测研究[J].现代电子技术,2017,40(4):12-15,4.基金项目
国家自然科学基金资助项目:异构多核并行机上线性代数方程组的快速算法研究 （）
