Abstract To improve the recognition accuracy and the convergence speed of the convolutional neural network algorithm, a convolutional neural network algorithm based on double optimization is proposed. By modeling a convolutional neural network and optimizing the process of feature extraction and regression classification, an optimization convolutional neural network is built. Thus, the integrated optimization of the convolution and the full-connection process is realized. Compared with the local optimization network, the integrated optimization network obtains a higher convergence speed and better recognition accuracies. The experiments are conducted based on handwritten digit datasets and face datasets and the results show the improvement of the convergence speed and the recognition accuracy. And the effectiveness of the proposed algorithm is demonstrated. Moreover, this optimization strategy can be further extended into other deep learning algorithms related to convolution neural networks.
About author:: LIU Wanjun, born in 1959, master, professor. His research interests include digital image processing, moving object detection and tracking.LIANG Xuejian(Corresponding author), born in 1993, master student. His research interests include deep learning, image detection and pattern recognition.QU Haicheng, born in 1981, Ph.D., associate professor. His research interests include hyperspectral remote sensing image and GPU parallel computing.)
LIU Wanjun,LIANG Xuejian,QU Haicheng. Convolutional Neural Network Algorithm Based on Double Optimization for Image Recognition[J]. , 2016, 29(9): 856-864.
[1] 谢智歌,王岳青,窦 勇,等.基于卷积-自动编码机的三维形状特征学习.计算机辅助设计与图形学学报, 2015, 27(11): 2058-2064. (XIE Z G, WANG Y Q, DOU Y, et al. 3D Feature Learning via Convolutional Auto-Encoder Extreme Learning Machine. Journal of Computer-Aided Design & Computer Graphics, 2015, 27(11): 2058-2064.) [2] HINTON G E, SALAKHUTDINOV R R. Reducing the Dimensio-nality of Data with Neural Networks. Science, 2006, 313(5786): 504-507. [3] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet Classification with Deep Convolutional Neural Networks [C/OL]. [2016-03-11]. http://machinelearning.wustl.edu/mlpapers/paper_files/NIPS2012_0534.pdf. [4] 吴 飞,朱文武,于俊清.多媒体技术研究: 2014-深度学习与媒体计算.中国图象图形学报, 2015, 20(11): 1423-1433. (WU F, ZHU W W, YU J Q. Researches on Multimedia Technology 2014-Deep Learning and Multimedia Computing. Journal of Image and Graphics, 2015, 20(11): 1423-1433.) [5] CHEN Y S, LIN Z H, ZHAO X, et al. Deep Learning-Based Cla-ssification of Hyperspectral Data. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(6): 2094-2107. [6] 胡 振,傅 昆,张长水.基于深度学习的作曲家分类问题.计算机研究与发展, 2014, 51(9): 1945-1954. (HU Z, FU K, ZHANG C S. Audio Classical Composer Identification by Deep Neural Network. Journal of Computer Research and Development, 2014, 51(9): 1945-1954.) [7] 曾智勇,刘侍刚.加权的多尺度多分辨率人脸描述与识别方法.模式识别与人工智能, 2014, 27(4): 378-384. (ZENG Z Y, LIU S G. Weighted Multiscale and Multiresolution Face Description and Recognition Method. Pattern Recognition and Artificial Intelligence, 2014, 27(4): 378-384.) [8] PANG S C, YU Z Z. Face Recognition: A Novel Deep Learning Approach. Journal of Optical Technology, 2015, 82(4): 237-245. [9] ZHANG C, ZHANG Z Y. Improving Multiview Face Detection with Multi-task Deep Convolutional Neural Networks // Proc of the IEEE Winter Conference on Applications of Computer Vision. New York, USA: IEEE, 2014: 1036-1041. [10] 程 帅,孙俊喜,曹永刚,等.多示例深度学习目标跟踪.电子与信息学报, 2015, 37(12): 2906-2912. (CHENG S, SUN J X, CAO Y G, et al. Target Tracking Based on Multiple Instance Deep Learning. Journal of Electronics & Information Technology, 2015, 37(12): 2906-2912.) [11] 黄凯奇,任伟强,谭铁牛.图像物体分类与检测算法综述.计算机学报, 2014, 37(6): 1225-1240. (HUANG K Q, REN W Q, TAN T N. A Review on Image Object Classification and Detection. Chinese Journal of Computers, 2014, 37(6): 1225-1240.) [12] SMIRNOV E A, TIMOSHENKO D M, ANDRIANOV S N. Comparison of Regularization Methods for ImageNet Classification with Deep Convolutional Neural Networks. AASRI Procedia, 2014, 6: 89-94. [13] 张 军,韦志辉.一种基于卷积积分的图像去噪变分方法.中国图象图形学报, 2008, 13(9): 1673-1678. (ZHANG J, WEI Z H. A Variation Method Based on Convolution Integral for Image Denoising. Journal of Image and Graphics, 2008, 13(9): 1673-1678.) [14] SYAFEEZA A R, KHALIL-HANI M, LIEW S S, et al. Convolutional Neural Network for Face Recognition with Pose and Illumination Variation. International Journal of Engineering and Technology, 2014, 6(1): 44-57. [15] 蔡体健,樊晓平,谢 昕,等.基于编码复杂度的混合结构稀疏人脸识别方法.模式识别与人工智能, 2015, 28(7): 595-602. (CAI T J, FAN X P, XIE X, et al. Face Recognition Method of Mixed Structured Sparsity Based on Coding Complexity. Pattern Recognition and Artificial Intelligence, 2015, 28(7): 595-602.) [16] JI S W, XU W, YANG M, et al. 3D Convolutional Neural Networks for Human Action Recognition. IEEE Trans on Pattern Ana-lysis and Machine Intelligence, 2013, 35(1): 221-231. [17] LEVI G, HASSNCER T. Age and Gender Classification Using Convolutional Neural Networks // Proc of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. New York, USA: IEEE, 2015: 34-42. [18] LI Q, CAI W D, WANG X G, et al. Medical Image Classification with Convolutional Neural Network // Proc of the 13th International Conference on Control Automation Robotics and Vision. New York, USA: IEEE, 2014: 844-848. [19] YUAN A Q, BAI G, YANG P, et al. Handwritten English Word Recognition Based on Convolutional Neural Networks // Proc of the International Conference on Frontiers in Handwriting Recognition. New York, USA: IEEE, 2012: 207-212. [20] CIRESAN D C, MEIER U, GAMBARDELLA L M, et al. Convolutional Neural Network Committees for Handwritten Character Classification // Proc of the International Conference on Document Analysis and Recognition. Washington, USA: IEEE, 2011: 1135-1139. [21] 赵志宏,杨绍普,马增强.基于卷积神经网络LeNet-5的车牌字符识别研究.系统仿真学报, 2010, 22(3): 638-641. (ZHAO Z H, YANG S P, MA Z Q. License Plate Character Re-cognition Based on Convolutional Neural Network LeNet-5. Journal of System Simulation, 2010, 22(3): 638-641.)
2016, Vol. 29 
