基于残差胶囊增强网络的手写化学方程式配平

doi:10.3969/j.issn.2095-2198.2023.01.014

摘要
参考文献
相关文章
推荐阅读
Metrics

下载:

PDF (1471KB)
输出: BibTeX | EndNote (RIS)

摘要

针对人工手动录入化学方程式至电子文档效率低、手动配平化学方程式高计算量、难度大的问题，提出一种残差胶囊增强网络（residual capsule enhanced network）自动识别和自动录入手写化学方程式，并且计算机自动配平方程式的新方法.为准确将化学方程式分割为单个化学字符，采用改进的投影法和最小外边框组合算法进行化学方程式切分；在传统胶囊网络的基础上，将残差卷积子网络和增强型胶囊进行结合，并引入两个级别的初级胶囊以提高网络获取多样性特征性能，进而解决化学方程式识别率低的问题；通过采用改进的高斯消元算法解决化学方程式自动配平问题.对比实验结果表明：残差胶囊增强网络识别率为94.97%，优于传统的经典算法.改进的高斯消元配平算法的准确率为95.35%，比传统的高斯消元配平算法提升了9.46%.该算法在手写化学方程式自动识别及配平方面具有一定的应用价值.

	服务
	把本文推荐给朋友
	加入引用管理器
	E-mail Alert
	RSS
	（）
	作者相关文章

关键词: 胶囊网络')" href="#">

胶囊网络胶囊增强算法残差卷积动态路由高斯消元算法

Abstract:

In view of the low efficiency of manual input of chemical equations into electronic documents,high computational complexity and difficulty in manual balancing chemical equations,a new method for automatic recognition and automatic input of handwritten chemical equations by residual capsule enhanced network and automatic square matching program by computer was proposed.In order to segment the chemical equation into a single chemical character accurately,the improved projection method and the minimum outer border combination algorithm are used to segment the chemical equation.On the basis of the traditional capsule network,the residual convolution sub network is combined with the enhanced capsule,and two levels of primary capsule are introduced to improve the performance of the network to obtain the diversity characteristics,and then the chemical equation recognition is solved.An improved Gaussian elimination algorithm is used to solve the problem of chemical equation automatic balancing.The experimental results show that the recognition rate of residual capsule enhanced network is 94.97%,which is better than the traditional classical algorithm.The accuracy of the improved Gaussian elimination algorithm is 95.35%,which is 9.46% higher than the traditional Gaussian elimination algorithm.This algorithm has certain application value in automatic recognition and balancing of handwritten chemical equations.

Key words: capsule network')" href="#">

capsule network capsule enhancement algorithm residual convolution dynamic routing Gaussian elimination algorithm

出版日期: 2023-02-27 发布日期: 2024-06-06 整期出版日期: 2023-02-27

ZTFLH:

TP391

基金资助:

辽宁省教育厅科学研究项目(LQ2017008)；辽宁省高等教育学会“十三五”规划高教研究项目(GHYB160163)；沈阳化工大学教育教学培育工程项目(2020 No.35)

引用本文:

李大舟, 张小明, 高巍.

基于残差胶囊增强网络的手写化学方程式配平 [J]. 沈阳化工大学学报, 2023, 37(1): 87-96.
LI Dazhou, ZHANG Xiaoming, GAO Wei.

The Realization of Capsule Balance Based on Handwritten Chemical Equation . Journal of Shenyang University of Chemical Technology, 2023, 37(1): 87-96.

链接本文:

https://xuebao.syuct.edu.cn/CN/10.3969/j.issn.2095-2198.2023.01.014 或 https://xuebao.syuct.edu.cn/CN/Y2023/V37/I1/87

［1］惠崇艳,周阳.浅谈初中化学方程式的配平［J］.明日,2019(4):79.

［2］缪徐.基于“透”字特征体现的概念教学创新——以初中“化学方程式”的教学为例［J］.化学教学,2019(5):61-64.

［3］袁德成,潘多涛,于三三,等.教学形成性评价的研究进展［J］.沈阳化工大学学报,〖JP〗2019,33(1):1-7.

［4］AKINOL R O,KUTCHIN S Y,NYAM I A,et al.Using Row Reduced Echelon Form in Balancing Chemical Equations［J］.Advances in Linear Algebra & Matrix Theory,2016,6(4):146-157.

［5］冯露.化学方程式的配平多解探究［J］.化学教学,2018(4):78-81.

［6］陆萍.初中化学方程式的配平方法探析［J］.华夏教师,2018(30):65.

［7］张天柱,邹承明.使用模糊聚类的胶囊网络在图像分类上的研究［J］.计算机科学,2019,46(12):279-285.

［8］YANG J F,WANG K,LI M D,et al.A Novel Method for Recognition of Handwritten Chemical Symbols Based on MQDF［C］//2011 4th International Congress on Image and Signal Processing.Shanghai：IEEE,2011:1490-1493.

［9］TANG P,HUI S C,FU C W.Online Chemical Symbol Recognition for Handwritten Chemical Expression Recognition［C］//2013 IEEE/ACIS 12th International Conference on Computer and Information Science（ICIS）.Niigata:IEEE,2013:535-540.

［10］DENG F,PU S L,CHEN X H,et al.Hyperspectral Image Classification with Capsule Network Using Limited Training Samples［J］.Sensors,2018,18(9):3153.

［11］XIANG C Q,ZHANG L,TANG Y,et al.MS-CapsNet:A Novel Multi-Scale Capsule Network［J］.IEEE Signal Processing Letters,2018,25(12):1850-1854.

［12］王志刚,夏慧明.基于差异演化算法的化学方程式配平研究［J］.哈尔滨商业大学学报(自然科学版),2012,28(4):455-458.

［13］陈素彬.利用量子粒子群优化算法配平化学方程式［J］.广州化工,2016,44(10):201-203.

［14］ZABADI A M,ASSAF R.From Chemistry to Linear Algebra:Balancing a Chemical Reaction Equation Using Algebraic Approach［J］.International Journal of Advanced Biotechnology and Research,2017,8(1):24-33.

［15］SABOUR S,FROSST N,HINTON G E.Dynamic Routing Between Capsules［C］//Proceedings of the 31st International Conference on Neural Information Processing Systems.Long Beach：Curran Associates Inc.,2017:3859-3869.

［16］LI X，HU X L，YANG J.SpatialGroup-Wise Enhance:Improving Semantic Feature Learning in Convolutional Networks［EB/OL］.(2019-05-25)［2020-07-20］.https://arxiv.org/pdf/1905.09646.pdf.

［17］HUANG G,LIU Z,VAN DER MAATEN L,et al.Densely Connected Convolutional Networks［C］//2017 IEEE Conference on Computer Vision and Pattern Recognition（CVPR）.Honolulu：IEEE,2017:2261-2269.

［18］DELIGE A，CIOPPA A，VAN DROOGENBROECK M.HitNet:A Neural Network with Capsules Embedded in a Hit-or-Miss Layer,Extended with Hybrid Data Augmentation and Ghost Capsules［EB/OL］.(2018-06-18)［2020-07-05］.https://arxiv.org/pdf/1806.06519.pdf.

[1]	孙丹, 朱立军. 基于BP神经网络和霍夫变换的非理想虹膜定位 [J]. 沈阳化工大学学报, 2023, 37(1): 60-67.
[2]	赵立杰, 吴志豪, 黄明忠. 基于SIFT特征匹配的活性污泥显微图像拼接方法 [J]. 沈阳化工大学学报, 2023, 37(1): 74-79.
[3]	高巍, 刘博洋, 李大舟, 王淮中. 基于LSTM-StackingCXR模型的房价预测算法研究 [J]. 沈阳化工大学学报, 2023, 37(1): 80-86.
[4]	张翔宇, 朱立军. 改进的Faster R-CNN海洋鱼类检测模型 [J]. 沈阳化工大学学报, 2022, 36(6): 562-568.
[5]	唐忠, 刘华阳, 刘珊珊, 李大舟. 基于认知概念的谓词扩展与表示 [J]. 沈阳化工大学学报, 2022, 36(6): 569-576.
[6]	张伟, 王倚天. 基于高频RFID的移动端图书自助借还系统设计与实现 [J]. 沈阳化工大学学报, 2022, 36(5): 476-480.

No Suggested Reading articles found!

Viewed

Full text

Abstract

Cited

Shared

Discussed