基于深度学习的盲文识别方法

doi:10.3969/j.issn.1006-2475.2015.06.008

摘要/Abstract

摘要： 提出一种基于深度学习的盲文点字识别方法，利用深度模型——堆叠去噪自动编码器（Stack Denoising AutoEncoder，SDAE）解决盲文识别中特征的自动提取与降维等问题。在构建深度模型过程中，采用非监督贪婪逐层训练算法（Greedy LayerWise Unsupervised Learning Algorithm）初始化网络权重，使用反向传播算法优化网络参数。利用SDAE自动学习盲文点字图片特征，使用Softmax分类器进行识别。实验结果表明，本文所提方法较之传统方法，可以有效解决样本特征的自动学习与特征降维等问题，操作更为简易，并能获得满意的识别结果。

关键词: 盲文识别, 深度学习, 特征提取, SDAE, 神经网络

Abstract: This paper mainly proposes a deep learning method, using Stacked Denoising AutoEncoder (SDAE) to solve the problems of automatic feature extraction and dimension reduction in Braille recognition. In the construction of a network with deep architecture, a feature extractor was trained with unsupervised greedy layerwise training algorithm to initialize the weights for extracting features from Braille images, and then a following classifier was set up for recognition. The experimental results show that comparing to traditional methods, the constructed network based on the deep learning method can easily recognize Braille images with satisfied performance. The deep learning model can effectively solve the Braille recognition problem in automatic feature extraction and dimension reduction with a reduced preprocessing.

Key words: Braille recognition, deep learning, feature extraction, SDAE, neural network

中图分类号:

TP391

李婷. 基于深度学习的盲文识别方法[J]. 计算机与现代化, 2015, 0(6): 37-40.

LI Ting. A Deep Learning Method for Braille Recognition[J]. Computer and Modernization, 2015, 0(6): 37-40.

参考文献

［1］张贵建. 盲人高等中医教育——中诊教程的改革与思考［EB/OL］. http://www.zgmx.org.cn/ViewInfo.asp?id=1958, 2007-04-02.
［2］ Zhang Shanjun, Yoshino Kazuyoushi. A braille recognition system by the mobile phone with embedded camera［C］// 2007 the 2rd International Conference Innovation Computing, Information and Control (ICICIC). Japan, 2007:223.
［3］尹佳,李杰,王丽荣. 盲文自动识别方法研究［J］. 长春大学学报, 2010,20(8):54-56.
［4］李念峰,董迎红,肖志国. 基于图像处理的盲文自动识别系统研究［J］. 制造业自动化, 2012,34(2):63-67.
［5］ Poirson P, Idrees H. Multimodal Stacked Denoising Autoencoders［EB/OL］. http://crcv.ucf.edu/REU/reu2013/patrick_poirson/report.pdf, 2014-11-05.
［6］ Baldi P. Autoencoders, unsupervised learning, and deep architectures［J］. JMLR: Workshop and Conference Proceedings. 2012,27:37-50.
［7］ Glüge S, Bck R, Wendemuth A. AutoEncoder PreTraining of SegmentedMemory recurrent neural networks［C］// Proceedings of the 21st European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning(ESANN 2013). 2013:29-34.
［8］ Lu Xugang, Tsao Y, Matsuda S, et al. Speech enhancement based on deep denoising autoencoder［C］// The 14th Annual Conference of the International Speech Communication Association. 2013:436440.
［9］ Vincent P. A connection between score matching and denoising autoencoders［J］. Neural Computation, 2011,23(7):16611674.
［10］Chen Minmin, Xu Zhixiang, Weinberger K, et al. Marginalized denoising autoencoders for domain adaptation［C］// Proceedings of the 29th International Conference on Machine Learning. 2012:1-8.
［11］Pascal Vincent, Hugo Larochelle, Yoshua Bengio, et al. Extracting and composing robust features with denoising autoencoders［C］// Proceedings of the 25th International Conference (ICML 2008) on Machine Learning. 2008:1096-1103.
［12］Lu Xugang, Tsao Y, Matsuda S, et al. Ensemble modeling of denoising autoencoder for speech spectrum restoration［C］// Proc. Interspeech 2014. 2014,14:885-889.
［13］Wu Daoxi, Pan Wei, Xie Lidang, et al. An adaptive stacked denoising autoencoder architecture for human action recognition［J］. Applied Mechanics and Materials. 2014,631-632:403-409.
［14］Sang Ruoxin, Jin Peiquan, Wan Shouhong. Discriminative feature learning for action recognition using a stacked denoising autoencoder［J］. Intelligent Data Analysis and its Applications, 2014,297:521-531.
［15］Vincent P, Larochelle H, Lajoie I, et al. Stacked denoising autoencoders: Learning useful representations in a deep network with a local denoising criterion［J］. Journal of Machine Learning Research, 2010,11:3371-3408.
［16］Zhu Ciyou, Byrd R H, Lu Peihuang, et al. Algorithm 778: LBFGSB:Fortran subroutines for largescale boundconstrained optimization［J］. ACM Transactions on Mathematical Software, 1997,23(4):550-560.
［17］Deng L, Seltzer M, Yu Dong, et al. Binary coding of speech spectrograms using a deep autoencoder［C］// Interspeech 2010. 2010:16921695.
［18］Lange S, Riedmiller M. Deep autoencoder neural networks in reinforcement learning［C］// International Joint Conference on Neural Networks. 2010:1-8.

[1]	何思达, 陈平华. 基于意图的轻量级自注意力序列推荐模型[J]. 计算机与现代化, 2024, 0(12): 1-9.
[2]	张晓东1, 白广芝1, 李敏1, 李昊洋2. 基于经验小波变换的油气井产量预测模型 [J]. 计算机与现代化, 2024, 0(12): 53-58.
[3]	刘云海1, 冯广1, 吴晓婷2, 杨群2. 复杂施工场景下的安全帽佩戴检测算法[J]. 计算机与现代化, 2024, 0(12): 66-71.
[4]	刘宝宝, 杨菁菁, 陶露, 王贺应. 基于注意力的DSMSC的遥感图像场景分类[J]. 计算机与现代化, 2024, 0(12): 72-77.
[5]	陈亮, 李诚, 易伟, 熊伟, 汪晓帆, 唐海东. 基于毫米波雷达与视觉融合的电力现场安全帽佩戴检测[J]. 计算机与现代化, 2024, 0(12): 100-107.
[6]	祁贤, 刘大铭, 常佳鑫. 基于改进自注意力机制的多视图三维重建[J]. 计算机与现代化, 2024, 0(11): 106-112.
[7]	陈凯1, 李宜汀1, 2, 全华凤1 . 基于改进YOLOv8的河道废弃瓶检测方法[J]. 计算机与现代化, 2024, 0(11): 113-120.
[8]	杨骏1, 胡为1, 朱文福2. 基于改进MobileNetV3的视觉SLAM回环检测算法[J]. 计算机与现代化, 2024, 0(10): 21-26.
[9]	王莹莹, 郝潇. 基于Res2Net和递归门控卷积的细粒度图像分类[J]. 计算机与现代化, 2024, 0(10): 74-79.
[10]	史星宇1, 李强2, 庄莉3, 梁懿3, 王秋琳3, 陈锴3, 伍臣周3, 常胜1. 一种面向工业部署的目标检测模型蒸馏技术[J]. 计算机与现代化, 2024, 0(10): 93-99.
[11]	马钰, 杨勇, 任鸽, 帕力旦·吐尔逊. 基于GCN和微调BERT的作文自动评分方法[J]. 计算机与现代化, 2024, 0(09): 33-37.
[12]	陈雪松1, 李衡1, 王浩畅2. 结合注意力机制和Mengzi模型的短文本分类[J]. 计算机与现代化, 2024, 0(09): 101-106.
[13]	张泽1, 张建权2, 3, 周国鹏2, 3. 基于改进YOLOv8s的摄像头模组缺陷检测[J]. 计算机与现代化, 2024, 0(09): 107-113.
[14]	余晨曦, 谷林. 基于人体骨架的电梯内异常行为识别预警[J]. 计算机与现代化, 2024, 0(09): 114-120.
[15]	程亚子1, 雷亮1, 2, 陈瀚1, 赵毅然1. 基于转置注意力的多尺度深度融合单目深度估计[J]. 计算机与现代化, 2024, 0(09): 121-126.