[1]谷文成,高谷九祥,凌卓毅,等.基于 PYNQ 的手写体数字识别系统设计实现[J].计算机技术与发展,2022,32(S2):31-35.[doi:10. 3969 / j. issn. 1673-629X. 2022. S2. 005]
　GU Wen-cheng,GAO Gu-jiu-xiang,LING Zhuo-yi,et al.Design and Implementation of Handwritten Digit Recognition System on Mobile Terminal Based on PYNQ[J].,2022,32(S2):31-35.[doi:10. 3969 / j. issn. 1673-629X. 2022. S2. 005]
点击复制

基于 PYNQ 的手写体数字识别系统设计实现()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
32
期数:
2022年S2期
页码:
31-35
栏目:
人工智能
出版日期:
2022-12-11

文章信息/Info

Title:
Design and Implementation of Handwritten Digit Recognition System on Mobile Terminal Based on PYNQ
文章编号:
1673-629X(2022)S2-0031-05
作者:
谷文成; 高谷九祥; 凌卓毅; 肖? 建
南京邮电大学 电子与光学工程学院、微电子学院,江苏 南京 210023
Author(s):
GU Wen-cheng; GAO Gu-jiu-xiang; LING Zhuo-yi; XIAO Jian
School of Electronic and Optical Engineering,Nanjing University of Posts andTelecommunications,Nanjing 210023,China
关键词:
手写体数字识别; 卷积神经网络; PYNQ; 软硬件协同设计; 移动端部署
Keywords:
handwritten digit recognition; convolutional neural network; PYNQ; software and hardware co - design; mobile terminal de鄄ployment
分类号:
TP391. 4
DOI:
10. 3969 / j. issn. 1673-629X. 2022. S2. 005
摘要:
手写体数字的识别作为图像识别领域的一项重要分支,广泛应用于银行汇款单号、人口普查、财务报表等大规模数据统计领域中。 而传统的手写体数字识别系统一般采用 CPU 或 GPU 的平台,有着功耗和成本较高、难以部署在移动端等弊端。 针对上述问题,设计了一种以卷积神经网络为基础,基于 PYNQ 的手写体数字识别系统。 通过软硬件协同设计的方式,合理划分软硬件任务来降低系统功耗。 首先在电脑端搭建卷积神经网络模型,通过训练验证,以获取权重和偏置等技术参数,并转换为相应的二进制格式文件。 之后在 VIVADO HLS 工具中设计完成了卷积层和最大池化层的 IP 核模块设计,以及系统的连线。 最后设计实现对应的上位机程序进行调控。 在 MNIST 数据集的测试下,该系统的识别准确率达到了 99. 07% ,功耗仅为 1. 54 W,相比于其他类似工作具有明显优势。
Abstract:
As an important branch in the field of image recognition,the recognition of handwritten digits is widely used in large-scale datastatistics fields such as bank remittance number, census, and financial statements. However, traditional handwritten digit recognitionsystems are generally based on CPU or GPU platforms, which have disadvantages such as high power consumption and cost,and difficultyin deploying on mobile terminals. In response to these problems,we design a PYNQ-based handwritten digit recognition system based onconvolutional neural networks. Through the idea of software and hardware co - design, tasks are divided reasonably. Firstly, theconvolutional neural network model is built on the PC side. Through training and verification,the technical parameters such as weight andbias are obtained and converted into the corresponding binary format file. Afterwards,the IP core modules of the convolutional layer andthe maximum pooling layer were designed in the VIVADO HLS tool,as well as the connection of the system. Finally,the correspondinghost computer program is designed and implemented for regulation. Under the test of the MNIST data set,the recognition accuracy of thesystem reached 99. 07% ,and the power consumption was only 1. 54 W,which has obvious advantages compared to other similar tasks.

相似文献/References:

[1]耿西伟 张猛 沈建京.基于结构特征分类BP网络的手写数字识别[J].计算机技术与发展,2007,(01):130.
　GENG Xi-wei,ZHANG Meng,SHEN Jian-jing.Recognition of Handwritten Numerals with Grouped BP Net Based on Structural Features[J].,2007,(S2):130.
[2]李琼,陈利,王维虎.基于SVM的手写体数字快速识别方法研究[J].计算机技术与发展,2014,24(02):205.
　LI Qiong[],CHEN Li[],WANG Wei-hu[].Research on Method of Fast Handwritten Digits Recognition Based on SVM[J].,2014,24(S2):205.
[3]崔凤焦.表情识别算法研究进展与性能比较[J].计算机技术与发展,2018,28(02):145.[doi:10．3969/j．issn．1673－629X．2018．02．031]
　CUI Feng-jiao.Ｒesearch and Performance Comparison of Facial Expression Ｒecognition Algorithm[J].,2018,28(S2):145.[doi:10．3969/j．issn．1673－629X．2018．02．031]
[4]张丹丹,李雷. 基于PCANet-RF的人脸检测系统[J].计算机技术与发展,2016,26(02):31.
　ZHANG Dan-dan,LI Lei. Face Detection System Based on PCANet-RF[J].,2016,26(S2):31.
[5]陈强锐,谢世朋.基于深度学习的肺部肿瘤检测方法[J].计算机技术与发展,2018,28(04):201.[doi:10.3969/ j. issn.1673-629X.2018.04.043]
　CHEN Qiang-rui,XIE Shi-peng.Lung Cancer Detection Method Based on Deep Learning[J].,2018,28(S2):201.[doi:10.3969/ j. issn.1673-629X.2018.04.043]
[6]郭子琰,舒心,刘常燕,等.基于ReLU 函数的卷积神经网络的花卉识别算法[J].计算机技术与发展,2018,28(05):154.[doi:10．3969/j．issn．1673－629X．2018．05．035]
　GUO Ziyan,SHU Xin,LIU Changyan,et al.A Recognition Algorithm of Flower Based on Convolution Neural Network with ReLU Function[J].,2018,28(S2):154.[doi:10．3969/j．issn．1673－629X．2018．05．035]
[7]缪宇杰,吴智钧,宫 婧.基于3D 卷积的视频错帧筛选方法[J].计算机技术与发展,2018,28(05):179.[doi:10.3969/ j. issn.1673-629X.2018.05.040]
　MIAO Yu-jie,WU Zhi-jun,GONG Jing.A Wrong Temporal-order Frames Identification Method Based on 3D Convolution[J].,2018,28(S2):179.[doi:10.3969/ j. issn.1673-629X.2018.05.040]
[8]吴玉枝,吴志红,熊运余.基于卷积神经网络的小样本车辆检测与识别[J].计算机技术与发展,2018,28(06):1.[doi:10.3969/ j. issn.1673-629X.2018.06.001]
　WU Yu-zhi,WU Zhi-hong,XIONG Yun-yu.Vehicle Detection and Recognition of a Few Samples Based on Convolutional Neural Network[J].,2018,28(S2):1.[doi:10.3969/ j. issn.1673-629X.2018.06.001]
[9]李相桥,李晨,田丽华,等.卷积神经网络并行训练的优化研究[J].计算机技术与发展,2018,28(08):12.[doi:10.3969/ j. issn.1673-629X.2018.08.003]
　LI Xiang-qiao,LI Chen,TIAN Li-hua,et al.Research on Optimization of Parallel Training for Convolution Neural Network[J].,2018,28(S2):12.[doi:10.3969/ j. issn.1673-629X.2018.08.003]
[10]邓宗平,赵启军,陈虎. 基于深度学习的人脸姿态分类方法[J].计算机技术与发展,2016,26(07):11.
　DEND Zong-ping,ZHAO Qi-jun,CHEN Hu. Face Pose Classification Method Based on Deep Learning[J].,2016,26(S2):11.

更新日期/Last Update: 2022-10-10