基于深度卷积神经网络的图像目标检测

被引:10
作者
尹勰
闫磊
机构
[1] 广东工业大学自动化学院
关键词
前景目标检测; 卷积神经网络; 反卷积; 金字塔池化;
D O I
暂无
中图分类号
TP391.41 []; TP183 [人工神经网络与计算];
学科分类号
080203 ; 081104 ; 0812 ; 0835 ; 1405 ;
摘要
针对从动态背景提取前景目标较差的问题,提出了一种基于卷积神经网络的图像前景目标检测方法。首先,基于传统的卷积神经网络构建了特征提取的网络模型,然后利用反卷积和金字塔池化,解决了传统卷积神经网络VGG-Net只能对整张图片分类以及只能输入固定尺寸图象的缺陷。针对R-CNN和SPP-Net网络模型提出了一种优化的boundingbox选择方法,使得对检测目标的定位更快更准确。在实际应用中,能够获得更好的前景目标检测效果,为后续的视频分析任务的研究提供了更好的条件。
引用
收藏
页码:96 / 97+100 +100
页数:3
相关论文
共 7 条
[1]  
"Multi-scale orderless pooling of deep convolutional activation features.". Gong,Yunchao,et al. Computer Vision–ECCV 2014 . 2014
[2]  
ImageNet Classification with Deep Convolutional Neural Networks. Krizhevsky A,Sutskever I,Hinton G E. Proceedings of 2012 Advances in Neural Information Processing Systems (NIPS 2012) . 2012
[3]  
Spatial pyramid pooling in deep convolutional networks for visual recognition. He K,Zhang X,Ren S, et al. Pattern Analysis and Machine Intelligence, IEEE Transactions on . 2015
[4]  
Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation. Girshick,Ross,Donahue,Jeff,Darrell,Trevor,et al. Computer Science . 2014
[5]  
Fully Convolutional Networks for Semantic Segmentation. Jonathan L,Evan S,Trevor D. IEEE Transactions on Pattern Analysis.&.Machin Intelligence . 2015
[6]  
Some Improvements on Deep Convolutional Neural Network Based Image Classification. Howard A G. Computer Science . 2013
[7]  
Visual Tracking with Fully Convolutional Networks. Wang L,Ouyang W,Wang X,et al. IEEE International Conference on Computer Vision . 2016