全部 标题 作者
关键词 摘要

OALib Journal期刊
ISSN: 2333-9721
费用:99美元

查看量下载量

相关文章

更多...

基于SVM的手写数字识别
Handwritten Digit Recognition Based on SVM

DOI: 10.12677/airr.2025.143049, PP. 501-509

Keywords: 支持向量机(SVM),手写数字识别,RBF核函数,Linear核函数
Support Vector Machines (SVM)
, Handwritten Digit Recognition, RBF Kernel Function, Linear Kernel Function

Full-Text   Cite this paper   Add to My Lib

Abstract:

本文主要研究基于支持向量机(SVM)的手写数字识别方法。通过对SVM算法的原理进行深入分析,结合手写数字图像的特征提取方法,构建了一个高效的手写数字识别系统。初始时,固定惩罚参数C (选用0.1)和γ值(选用scale),采用RBF核和linear核函数进行训练,最后对RBF核函数和linear核函数的各项性能进行比较,如召回率、准确率等,选取针对手写数字识别这一场景下最优的核函数。最后实验结果表明,RBF核函数的各项性能指标如准确率、召回率、F1函数等值在这种场景下要优于linear核函数。最后选取RBF核函数,再进一步调整惩罚参数Cγ值以训练出最优的模型。
This paper mainly studies the handwritten numeral recognition method based on Support Vector Machine (SVM). Through the in-depth analysis of the principle of SVM algorithm, combined with the feature extraction method of handwritten digital images, an efficient handwritten digital recognition system is constructed. At the beginning, the penalty parameters C (0.1) and γ (scale) are fixed, and RBF kernel and linear kernel function are used for training. Finally, the performances of RBF kernel function and linear kernel function, such as recall rate and accuracy rate, are compared, and the optimal kernel function for handwritten numeral recognition is selected. Finally, the experimental results show that the performance indexes of RBF kernel function, such as accuracy, recall and F1 function equivalence, are better than linear kernel function in this scenario. Finally, RBF kernel function is selected, and then the penalty parameters C and γ are further adjusted to train the optimal model.

References

[1]  田春婷. 基于PCA降维的MNIST手写数字识别优化[J]. 现代信息科技, 2024, 8(16): 64-68.
[2]  徐平, 徐海东, 杨拓, 等. 三层衍射神经网络实现手写数字识别[J]. 物理学报, 2022, 71(18): 137-144.
[3]  叶佩, 江涛. 基于BP特征筛选的模糊聚类手写体数字识别[J]. 武汉理工大学学报, 2007, 29(6): 128-130.
[4]  苗夺谦, 张红云, 李道国, 等. 基于主曲线的脱机手写数字识别[J]. 电子学报, 2005, 33(9): 1639-1643.
[5]  林列, 常胜江. 用于手写体数字不变性特征提取及识别的自组织算法[J]. 电子学报, 2003, 31(10): 1506-1509.
[6]  房孝猛, 王华来, 徐晖, 等. 基于SVM与近红外TDLAS技术的多组分痕量气体识别与检测[J]. 光谱学与光谱分析, 2024, 44(10): 2909-2915.
[7]  张微微, 璩怡, 王强, 等. 同步荧光光谱技术结合支持向量机对掺杂牛奶智能判别研究[J]. 光谱学与光谱分析, 2024, 44(9): 2428-2433.
[8]  龙泽凯, 陈聪飞, 郝东来. 基于CNN的手势识别[J]. 电脑编程技巧与维护, 2023(5): 104-106, 122.
[9]  侯礼灏, 迟玉红, 谢虹铭. 基于知识图谱的法律问答系统研究[J]. 电脑编程技巧与维护, 2023(8): 106-108.
[10]  Devi, G. and Madhavi, D. (2024) Detection and Classification of Diabetic Retinopathy through Identification of Blood Vessel Thickness Using FOFF& ML Classifiers. Journal of Harbin Institute of Technology (New Series), 6, 84-96.
https://kns.cnki.net/kcms/detail/23.1378.T.20241018.1036.002.html
[11]  Jose, A., Amrutha, S., Jain, J. and Joseena, S. (2021) A Study to Assess the Effect of Workshop on Knowledge Regarding Learning Disabilities among Nurses Working in Selected Institutions at Kottayam. International Journal of Advances in Nursing Management, 9, 46-48.
https://doi.org/10.5958/2454-2652.2021.00013.5
[12]  Plc, F.G.S. (2018) SVM GLOBAL FUND PLC—ANNUAL REPORT—SEPTEMBER 30, 2017.
[13]  Plc, F.G.S. (2017) SVM GLOBAL FUND PLC—ANNUAL REPORT—SEPTEMBER 30, 2016.
[14]  Thomas, W. and Singh, V. (2024) Accelerating Neural Network Inference in Handwritten Digit Recognition—Comparative Study. Parallel Processing Letters, 34, Article ID: 2430001.
https://doi.org/10.1142/s0129626424300010
[15]  Lin, Z., Li, J., Dai, G., Chen, T., Huang, S. and Lin, J. (2024) Contrastive Representation Enhancement and Learning for Handwritten Mathematical Expression Recognition. Pattern Recognition Letters, 186, 14-20.
https://doi.org/10.1016/j.patrec.2024.08.021
[16]  Xu, H. and Zhang, X. (2024) Recognizing Digital Ink Chinese Characters Written by International Students Using a Residual Network with 1-Dimensional Dilated Convolution. Information, 15, Article 531.
https://doi.org/10.3390/info15090531
[17]  Abdo, H.A., Abdu, A., Al-Antari, M.A., Manza, R.R., Talo, M., Gu, Y.H., et al. (2024) End-to-End Deep Learning Framework for Arabic Handwritten Legal Amount Recognition and Digital Courtesy Conversion. Mathematics, 12, Article 2256.
https://doi.org/10.3390/math12142256

Full-Text

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133