[1]王 林,董 楠.基于Gabor特征与卷积神经网络的人体轮廓提取[J].南京理工大学学报(自然科学版),2018,42(01):89.[doi:10.14177/j.cnki.32-1397n.2018.42.01.013]
 Wang Lin,Dong Nan.Human silhouette identification based on Gabor featureand convolutional neural network[J].Journal of Nanjing University of Science and Technology,2018,42(01):89.[doi:10.14177/j.cnki.32-1397n.2018.42.01.013]
点击复制

基于Gabor特征与卷积神经网络的人体轮廓提取()
分享到:

《南京理工大学学报》(自然科学版)[ISSN:1005-9830/CN:32-1397/N]

卷:
42卷
期数:
2018年01期
页码:
89
栏目:
出版日期:
2018-02-28

文章信息/Info

Title:
Human silhouette identification based on Gabor feature and convolutional neural network
文章编号:
1005-9830(2018)01-0089-07
作者:
王 林董 楠
西安理工大学 自动化与信息工程学院,陕西 西安 710048
Author(s):
Wang LinDong Nan
School of Automation and Information Engineering,Xi’an University of Technology,Xi’an 710048,China
关键词:
人体轮廓提取 Gabor特征 卷积神经网络 图像处理 深度学习 VOC2012数据集
Keywords:
human silhouette identification Gabor feature convolutional neural network picture processing deep learning VOC2012 data set
分类号:
TP391
DOI:
10.14177/j.cnki.32-1397n.2018.42.01.013
摘要:
为解决传统方法无法精确分割单幅图像中人体轮廓的问题,提出一种将Gabor小波特征与卷积神经网络结合的人体轮廓提取方法。首先计算人体图像8个方向的Gabor特征,并将计算结果传入改进的卷积神经网络进行人物模型训练; 再将待测图像计算Gabor特征后传入人物模型进行检测,从而输出人物掩膜; 对掩膜进行形态学处理并同原图像进行相与操作,最终得到人体轮廓。经VOC2012数据集上测试,该人体轮廓提取方法的准确度高达96%。实验结果表明,通过将传统特征提取方法与深度学习的特征学习方法相结合,不但提高了特征学习的速度还提高了检测的准确度。
Abstract:
Traditional methods can not divide the human silhouette accurately in a single image. Aiming at this problem,a method of human silhouette extraction based on Gabor feature and convolutional neural network(CNN)is proposed. Firstly the Gabor wavelet feature of human image in 8 directions is calculated and introduced into the improved CNN for the character model training; secondly the being-tested image is introduced into character model for detection after calculating Gabor wavelet feature,and the mask of the characters is output; finally the mask is processed by morphology and an operation is made with the original image to get the human silhouette. The detection accuracy of this method is up to 96% in the test of VOC2012 data set. The experimental results show that the combination of traditional feature extraction method and the feature learning method of deep learning improves the speed of feature learning and the detection accuracy.

参考文献/References:

[1] Cai Z,Saberian M,Vasconcelos N. Learning complexity-aware cascades for deep pedestrian detection[C]//IEEE International Conference on Computer Vision. Santiago,Chile:IEEE,2015:3361-3369. [2]Oren M,Papageorgiou C,Sinha P,et al. Pedestrian detection using wavelet templates[C]//Conference on Computer Vision and Pattern Recognition. San Juan,Puerto Rico,USA:IEEE Computer Society,1997:193. [3]Hosang J,Benenson R,Dollar P,et al. What makes for effective detection proposals?[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence,2016,38(4):814-830. [4]Dalal N,Triggs B. Histograms of oriented gradients for human detection[C]//IEEE Computer Society Conference on Computer Vision and Pattern Recognition,2005. San Diego,CA,USA:IEEE,2005:886-893. [5]Zhang S,Bauckhage C,Cremers A B. Informed Haar-like features improve pedestrian detection[C]//Computer Vision and Pattern Recognition. Columbus,OH,USA:IEEE,2014:947-954. [6]Dollár P,Tu Z,Perona P,et al. Integral channel features[C]//BMVC 2009:Proceedings of British Machine Vision Conference. London,UK:BMVA Press,2009:1-11. [7]Liu C,Wechsler H. Gabor feature based classification using the enhanced fisher linear discriminant model for face recognition[J]. IEEE Transactions on Image Processing,2002,11(4):467-475. [8]Maji S,Berg A C,Malik J. Classification using intersection kernel support vector machines is efficient[C]//IEEE Conference on Computer Vision and Pattern Recognition,2008. Anchorage,AK,USA:IEEE,2008:1-8. [9]FreundY,Schapire R E. A decision-theoretic generalization of on-line learning and an application to boosting[C]//Computational Learning Theory. Berlin:Springer Berlin Heidelberg,1995:119-139. [10]Chappelow J. Snakes:active contour models[J]. International Journal of Computer Vision,1988,1(4):321-331. [11]曹之江,郝矿荣,丁永生. 基于GVF-Snake人体轮廓提取的优化算法[J]. 计算机工程,2008,34(22):204-206. Cao Z J,Hao K R,Ding Y S. Optimization algorithm based on GVF-snake for body contour extraction[J]. Computer Engineering,2008,34(22):204-206. [12]Liu T,Yuan Z,Sun J,et al. Learning to detect a salient object.[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence,2010,33(2):353-367. [13]逯暄,肖泽龙,胡泰洋,等. 基于视觉显著性分析的毫米波辐射图像增强[J]. 南京理工大学学报,2014,38(1):134-139. Lu Xuan,Xiao Zelong,Hu Taiyang,et al. MMW radiometric image enhancement based on visual saliency analysis[J]. Journal of Nanjing University of Science and Technology,2014,38(1):134-139. [14]Hinton G,Salakhutdinov R. Reducing the dimensionality of data with neural networks[J].Science,2006,313:504-507. [15]Redmon J,Angelova A. Real-time grasp detection using convolutional neural networks[C]//IEEE International Conference on Robotics and Automation. Piscotaway,USA:IEEE,2015:1316-1322. [16]Long J,Shelhamer E,Darrell T. Fully convolutional networks for semantic segmentation[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence,2014,39(4):640-651. [17]姚伟,孙正兴,张岩. 面向脸部表情识别的Gabor特征选择方法[J]. 计算机辅助设计与图形学学报,2008,20(1):79-84. Yao W,Sun Z X,Zhang Y. Optimal Gabor feature for facial expression recognition[J]. Journal of Computer-Aided Design & Computer Graphics,2008,20(1):79-84. [18]Arivazhagan S,Ganesan L,Priyal S P. Texture classification using Gabor wavelets based rotation invariant features[M]. New York,NY,USA:Elsevier Science Inc,2006. [19]Krizhevsky A,Sutskever I,Hinton G E. ImageNet classification with deep convolutional neural networks[C]//International Conference on Neural Information Processing Systems. New York,USA:Curran Associates Inc,2012:1097-1105. [20]Szegedy C,Liu W,Jia Y,et al. Going deeper with convolutions[C]//Computer Vision and Pattern Recognition. Boston,MA,USA:IEEE,2015:1-9. [21]Simonyan K,Zisserman A. Very deep convolutional networks for large-scale image recognition[J]. arXiv preprint arXiv:1409.1556,2014.

备注/Memo

备注/Memo:
收稿日期:2017-11-20 修回日期:2017-12-25 基金项目:陕西省科技计划重点项目(2017ZDCXL-GY-05-03) 作者简介:王林(1963-),男,博士,教授,主要研究方向:深度学习、社交网络、数据挖掘等,E-mail:wanglin@xaut.edu.cn; 通讯作者:董楠(1993-),女,硕士生,主要研究方向:计算机视觉、图像处理,E-mail:dn_dora@163.com。 引文格式:王林,董楠. 基于Gabor特征与卷积神经网络的人体轮廓提取[J]. 南京理工大学学报,2018,42(1):89-95. 投稿网址:http://zrxuebao.njust.edu.cn
更新日期/Last Update: 2018-02-28