基于HOG的目标分类特征深度学习模型

doi:10.3969/j.issn.1000-3428.2016.12.031

计算机工程

基于HOG的目标分类特征深度学习模型

何希平^a,b,c,张琼华^d,刘波 ^a,b,c

(重庆工商大学 a.电子商务及供应链系统重庆市重点实验室;b.重庆市检测控制集成系统工程实验室; c.计算机科学与信息工程学院; d.图书馆,重庆 400067)

收稿日期:2015-10-19 出版日期:2016-12-15 发布日期:2016-12-15
作者简介:何希平(1968—),男,教授、博士,主研方向为计算机视觉、机器学习、信息安全;张琼华,馆员;刘波,讲师、博士。
基金项目:
重庆市教委科学技术研究计划项目(KJ1400612)。

Deep Learning Model of Object Classification Feature Based on HOG

HE Xiping^a,b,c,ZHANG Qionghua ^d,LIU Bo ^a,b,c

(a.Chongqing Key Laboratory of Electronic Commerce and Supply Chain System;b.Chongqing Engineering Laboratory for Detection,Control and Integrated System;c.School of Computer Science and Information Engineering; d.Library,Chongqing Technology and Business University,Chongqing 400067,China)

Received:2015-10-19 Online:2016-12-15 Published:2016-12-15

摘要/Abstract

摘要： 为提高低配置计算环境中的视觉目标实时在线分类特征提取的时效性和分类准确率,提出一种新的目标分类特征深度学习模型。根据高时效性要求,选用分类器模型离线深度学习的策略,以节约在线训练时间。针对网络深度受限和高识别率要求,提取图像的局部方向梯度直方图(HOG)特征,构建稀疏自编码器栈对HOG特征进行深层次编码,设计Softmax多分类器对所抽取的特征进行分类。在深度神经网络模型学习过程中,引入最小化各层结构风险和微调全网参数的二阶段最优化策略。利用场景图像库Caltech101和手写数字库MNIST的训练样本与测试样本进行对比实验,结果表明,该模型在局部特征提取方面的时效优于单层卷积神经网络(CNN)模型,分类准确率高于CNN、栈式自编码器等对比模型。

关键词: 计算机视觉, 目标分类, 方向梯度直方图特征, 栈式自编码器, 深度学习

Abstract: To improve the feature extraction timeliness and classification validity of real-time classification of visual object in low computing profile,a Histogram of Oriented Gradients(HOG)-based feature deep learning model for object classification is proposed.For the requirements of high timeliness,offline deep learning strategy is applied to the classifier model to save its online training time.In view of the requirements of network depth limitation and high recognition rate,the local feature of HOG feature of an image is extracted to be used as the input of the sparse autoencoder stack so as to output the high level feature code of the sample image.The Softmax multiple classifier is designed to classify the extracted features.During the learning process of the deep neural network model,the two-stage optimization strategy is introduced,which minimizes the structural risk of every layer and fine-tune the parameters of the whole model.Using some samples of the scene image database Caltech101 and that of the handwritten digits database MNIST as the training set and the others as the test set to perform the comparative experiment,results show that the time performance of the proposed model is better than that of one-layer only Convolutional Neural Network(CNN),and the classification accuracy of the trained model is higher than that of CNN,Stacked Autoencoder(SAE) comparative models.

Key words: computer vision, object classification, Histogram of Oriented Gradients(HOG) feature, Stacked Autoencoder(SAE), deep learning

中图分类号:

TP391

何希平,张琼华,刘波. 基于HOG的目标分类特征深度学习模型[J]. 计算机工程, doi: 10.3969/j.issn.1000-3428.2016.12.031.

HE Xiping,ZHANG Qionghua,LIU Bo. Deep Learning Model of Object Classification Feature Based on HOG[J]. Computer Engineering, doi: 10.3969/j.issn.1000-3428.2016.12.031.

参考文献

参考文献［1］Hinton G E,Osindero S,Teh Y.A Fast Learning Algorithm for Deep Belief Nets［J］.Neural Com-putation,2006,18(1):1527-1554. ［2］Hinton G E,Salakhutdinov R R.Reducing the Dimensionality of Data with Neural Networks［J］.Science,2006,313(5786):504-507. ［3］潘泓,金立左,夏思宇,等.基于多层次互补特征的通用目标检测模型［J］.电子与信息学报,2012,34(7):1531-1537. ［4］Socher R,Pennington J,Huang E,et al.Manning.Semi-supervised Recursive Autoencoders for Predicting Sentiment Distributions［C］//Proceedings of 2011 Con-ference on Empirical Methods in Natural Language Processing.Edinburgh,UK:［s.n. ］,2011:151-161. ［5］LeCun Y,Boser B,Denker J S,et al.Backpropagation Applied to Handwritten Zip Code Recognition［J］.Neural Computation,1989,1(4):541-551. ［6］LeCun Y L,Boser B J.Denker,et al.Handwritten Digit Recognition with a Backpropagation Neural Net-work［J］.Advances in Neural Information Processing Systems,1990,2(1):299-304. ［7］Lowe D G.Object Recognition from Local Scale-invariant Features［EB/OL］.(1999-09-21).http://dx.doi.org/10.1109/iccv. ［8］Lowe D G.Distinctive Image Features from Scale-invariant Keypoints［J］.International Journal of Computer Vision,2004,60(2):91-110. (下转第187页) (上接第180页) ［9］Sánchez J,Perronnin F,Mensink T,et al.Image Classification with the Fisher Vector:Theory and Practice［J］.International Journal of Computer Vision,2013,105(3):222-245. ［10］Sydorov V,Sakurada M,Lampert C H.Deep Fisher Kernels——End to End Learning of the Fisher Kernel GMM Parameters［C］//Proceedings of IEEE Con-ference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2014:1402-1409. ［11］Sánchez J,Redolfi J.Exponential Family Fisher Vector for Image Classification［J］.Pattern Recognition Letters,2015,59(7):26-32. ［12］Dalal N,Triggs B.Histograms of Oriented Gradients for Human Detection［C］//Proceedings of IEEE Conference on Computer Vision & Pattern Recognition.Washington D.C.,USA:IEEE Press,2005:886-893. ［13］Ma Ding,Pang Xin,Qu Mingzhi,et al.A Target Tracking Method Combining HOG and Shape Con-text［J］.Journal of Computational Information Systems,2014,10(8):7053-7060. ［14］王知音,禹龙,田生伟,等.基于栈式自编码的水体提取方法［J］.计算机应用,2015,35(9):2706-2709. ［15］Bengio Y,Lamblin P,Popovici P,et al.Greedy Layer-wise Training of Deep Networks［J］.Advances in Neural Information Processing Systems,2007,19(1):153-160. ［16］李海峰,李纯果.深度学习结构和算法比较分析［J］.河北大学学报(自然科学版),2012,32(5):538-544. ［17］Li Feifei,Fergus R,Perona P.Learning Generative Visual Models from Few Training Examples:An Incremental Bayesian Approach Tested on 101 Object Categories［J］.Computer Vision & Image Under-standing,2007,106(1):59-70. ［18］Yann Lecun,Cortes C,Christopher J C.The MNIST Ddatabase of Handwritten Digits［EB/OL］.(2013-10-30).http://yann.lecun.com/exdb/mnist/. 编辑索书志

[1]	姜妍, 张立国. 面向深度学习模型的对抗攻击与防御方法综述[J]. 计算机工程, 2021, 47(1): 1-11.
[2]	韩虎, 赵启涛, 孙天岳, 刘国利. 面向社交媒体评论的上下文语境讽刺检测模型[J]. 计算机工程, 2021, 47(1): 66-71.
[3]	刘瑄, 池明旻. 基于多粒度特征蒸馏的遥感图像场景分类研究[J]. 计算机工程, 2021, 47(1): 224-229,238.
[4]	徐龙壮, 彭力, 朱凤增. 多任务金字塔重叠匹配的行人重识别方法[J]. 计算机工程, 2021, 47(1): 239-245,254.
[5]	王润涵, 李兵, 滕奇志. 基于卷积神经网络的岩心FIB-SEM图像分割算法[J]. 计算机工程, 2021, 47(1): 264-274.
[6]	王嘉庆, 梅礼晔, 张俊华. 基于深度学习的手骨X射线图像骨龄评估[J]. 计算机工程, 2021, 47(1): 291-297.
[7]	李兰, 郑雨薇, 魏少玮, 胡克勇. 基于注意力机制的商家招牌分类研究[J]. 计算机工程, 2021, 47(1): 305-311.
[8]	肖经纬, 田军委, 王沁, 程希希, 王佳. 基于改进残差网络的果实病害分类方法[J]. 计算机工程, 2020, 46(9): 221-225.
[9]	生龙, 马建飞, 杨瑞欣, 吴迪. 基于特征交换的CNN图像分类算法研究[J]. 计算机工程, 2020, 46(9): 268-273.
[10]	辛伟瑶, 李剑, 王小亮, 李禹剑. 基于深度学习的地下浅层震源定位方法[J]. 计算机工程, 2020, 46(9): 292-297.
[11]	尚迪雅, 孙华, 洪振厚, 曾庆亮. 基于无梯度进化的神经架构搜索算法研究综述[J]. 计算机工程, 2020, 46(9): 16-26.
[12]	张国令, 王晓丹, 李睿, 来杰, 向前. 基于栈式降噪稀疏自编码器的极限学习机[J]. 计算机工程, 2020, 46(9): 61-67.
[13]	谢瑞, 邵堃, 霍星, MITHUN Md Masud Parvej. 一种改进的DIQaM_FR/NR图像质量评价模型[J]. 计算机工程, 2020, 46(8): 258-263,270.
[14]	赵亚南, 刘渊, 宋设. 融合多头自注意力机制的金融新闻极性分析[J]. 计算机工程, 2020, 46(8): 85-92.
[15]	熊亚辉, 陈东方, 王晓峰. 基于多尺度反向投影的图像超分辨率重建算法[J]. 计算机工程, 2020, 46(7): 251-259.

基于HOG的目标分类特征深度学习模型

Deep Learning Model of Object Classification Feature Based on HOG

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

Metrics

本文评价

推荐阅读 10