基于深度卷积神经网络的公式重复检测方法

doi:10.11896/jsjkx.200100108

摘要/Abstract

摘要： 近年来,随着教育智能化的发展,互联网教育模式成为了教育教学的重要载体。各类在线教育系统拥有海量试题资源,为学习者提供了便捷的学习途径。然而,试题来源繁多、收集方式不统一等因素,使得互联网中所积累的试题资源存在重复率高、质量较低的现象。因此,准确、高效地监测试题,是精炼网络资源、提高网络试题质量的重要方式。在这样的背景下,文中着重研究了针对理科试题资源中图片公式的重复检测问题,通过精准的公式识别检测,能够排除试题语义的干扰,进而加强试题资源监测。传统的公式重复检测方法,往往因为基于人工定义的各类规则,识别步骤繁琐,准确率和效率较低,难以应用于大规模的公式数据检测。据此,提出一种基于深度卷积神经网络的公式重复检测方法。首先,使用一种多通道卷积机制实现了公式图片特征提取和处理的自动化,使之适用于大规模的公式数据检测。然后,使用端到端的输出模式,避免了传统方法中间步骤过多可能导致误差累计的弊端。最后,为了验证模型的准确率以及实用性,在标准测试数据集以及模拟扫描图噪声的数据集上进行了充分的实验,实验结果表明此方法能够有效处理不同质量的公式图片,在检测精度和效率上取得了良好的结果。

关键词: 公式重复检测, 卷积神经网络, 试题质量, 图片识别

Abstract: In recent years,with the development of educational intelligence,the Internet education model has become an important carrier of education and teaching.Various online education systems provide learners with a convenient way to learn their vast amount of test resources.However,the accumulated exercise resources suffer from the high repetition rate and low quality due to various sources of test questions and inconsistent collection methods.Therefore,how to accurately and efficiently monitor test questions is an important way to refine network resources and improve the quality of network test questions.In this context,this paper focuses on the problem of repeated detection of picture formulas in science test resources.Through accurate formula recognition detection,it can eliminate the interference of test questions semantics,and then improve the test resource monitoring.In response to this problem,the traditional formula repeat detection method is often based on manually defined rules and difficult to apply to large-scale formula data detection because of cumbersome identification steps,low accuracy and low efficiency.Based on this,this paper proposes a formula repeated detection method based on deep convolutional neural network.Firstly,a multi-channel convolution mechanism is used to automate the extraction and processing of formula picture features,making it suitable for large-scale formula data detection.Then,using the end-to-end output mode,the accumulation of errors that may be caused by too many intermediate steps in the traditional method is avoided.Finally,in order to verify the accuracy and practicability of the model,this paper has carried out sufficient experiments on the standard test data set and the data set of the simulated scan noise.The experimental results show that this method can effectively process the formula pictures of different quality.Good results in both accuracy and efficiency.

Key words: Convolutional neural network, Duplicate formula detection, Exercise quality, Image recognition

中图分类号:

TP301

陈昂, 佟威, 周宇强, 阴钰, 刘淇. 基于深度卷积神经网络的公式重复检测方法[J]. 计算机科学, 2020, 47(11A): 409-415. https://doi.org/10.11896/jsjkx.200100108

CHEN Ang, TONG Wei, ZHOU Yu-qiang, YIN Yu, LIU Qi. Duplicate Formula Detection Based on Deep Convolutional Neural Network[J]. Computer Science, 2020, 47(11A): 409-415. https://doi.org/10.11896/jsjkx.200100108

参考文献

[1] BRESLOW L,PRITCHARD D E,DEBOER J,et al.Studying Learning in the Worldwide Classroom Research into edX's First MOOC[J].Research & Practice in Assessment,2013,8:13-25.
[2] POLSON M C.Foundations of Intelligent Tutoring Systems[M].Hove,UK:Psychology Press,2013.
[3] HUANG Z,LIU Q,CHEN E,et al.Question Difficulty Prediction for READING Problems in Standard Tests[C]//AAAI.2017:1352-1359.
[4] LIU Q,CHEN E H,ZHU T Y,et al.Research on Educational Data Mining for Online Intelligent Learning[J].Pattern Recognition and Artificial Intelligence,2018,31(1):77-90.
[5] KOHLHASE M,SUCAN I.A search engine for mathematicalformulae[C]//Proceedings of the 8th international conference on Artificial Intelligence and Symbolic Computation (AISC'06).Berlin:Springer-Verlag,2006:241-253.
[6] JADERBERG M,SIMONYAN K,VEDALDI A,et al.Reading Text in the Wild with Convolutional Neural Networks[J].ar-Xiv:1412.1842v1.
[7] LIN X Y,GAO L C,TANG Z.Mathematical Formula Identification and Performance EvaluationinPDFDocuments[J].International JournalonDocument Analysis and Recognition,2014,17(3):239-255.
[8] YIN Y,HUANG Z,CHEN E,et al.Transcribing Content from Structural Images with Spotlight Mechanism[C]//Proceedings of the 24th ACM SIGKDD International Conference on Know-ledge Discovery & Data Mining.ACM,2018:2643-2652.
[9] LIU Q,HUANG Z,HUANG Z,et al.Finding Similar Exercises in Online Education Systems[C]//Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.ACM,2018:1821-1830.
[10] WANG H,XU T,LIU Q,et al.MCNE:An End-to-End Framework for Learning Multiple Conditional Network Representations of Social Network[C]//Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD '19).Association for Computing Machinery,New York,NY,USA,2019:1064-1072.
[11] LIN X Y,GAO L C,TANG Z.A Text Line Detection Method for Mathematical Formula Recognition[C]//Proceedings of International Conference on Document Analysis and Recognition.2013:339-343.
[12] LI Y H,WANG K J,SHANG G W,et al.Baseline structure analysis and recognition algorithm research of mathematical formula[J].Computer Engineering and Applications,2008,44(16):18-22.
[13] ZANIBBI R.Recognition of mathematics notation via computer using baseline structure[R].Queen's University,Kingston,Ontario,2000.
[14] GUO J N.Research on Detection Algorithm of MathematicialFormula for MathML[D].Jinzhou:Bohai University,2016.
[15] ZHU H,NIE Z,DING M.Image recognition by affine moment invariants in Hartley transform domains[C]//International Symposium on Communications and Information Technologies.IEEE,2010:630-633.
[16] LI J,CHENG J,SHI J,et al.Brief Introduction of Back Propagation (BP) Neural Network Algorithmand Its Improvement[C]//Advances in Computer Science and Information Enginee-ring.Berlin:Springer.2012.
[17] LECUN Y,BENGIO Y.Convolutional networks for images,speech,and time series[J].The Handbook of Brain Theory and Neural Networks,1995,3361(10):1995.
[18] CHOPRA S,HADSELL R,LECUN Y.Learning a similaritymetric discrim-inatively,with application to face verification[C]//IEEE Computer Society Conference on Computer Vision and Pattern Recognition(CVPR 2005).IEEE,2005:539-546.
[19] ZAGORUYKO S,KOMODAKIS N.Learning to compare image patches via convolutional neural networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2015:4353-4361.
[20] IOFFE S,SZEGEDY C.Batch normalization:Accelerating deep network training by reducing internalcovariate shift[C]//International Conference on Machine Learning.2015:448-456.
[21] IOFFE S,SZEGEDY C.Batch normalization:Accelerating deep network training by reducing internal covariate shift[J].arXiv:1502.03167,2015.
[22] GLOROT X,BENGIO Y.Understanding the difficulty of training deep feedforward neural networks[C]//Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics.2010:249-256.
[23] HE K,ZHANG X,REN S,et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2016:770-778.
[24] HE K,ZHANG X,REN S,et al.Identity mappings in deep residual networks[C]//European Conference on Computer Vision.Cham:Springer,2016:630-645.

相关文章 15

[1]	周乐员, 张剑华, 袁甜甜, 陈胜勇. 多层注意力机制融合的序列到序列中国连续手语识别和翻译 Sequence-to-Sequence Chinese Continuous Sign Language Recognition and Translation with Multi- layer Attention Mechanism Fusion 计算机科学, 2022, 49(9): 155-161. https://doi.org/10.11896/jsjkx.210800026
[2]	李宗民, 张玉鹏, 刘玉杰, 李华. 基于可变形图卷积的点云表征学习 Deformable Graph Convolutional Networks Based Point Cloud Representation Learning 计算机科学, 2022, 49(8): 273-278. https://doi.org/10.11896/jsjkx.210900023
[3]	陈泳全, 姜瑛. 基于卷积神经网络的APP用户行为分析方法 Analysis Method of APP User Behavior Based on Convolutional Neural Network 计算机科学, 2022, 49(8): 78-85. https://doi.org/10.11896/jsjkx.210700121
[4]	朱承璋, 黄嘉儿, 肖亚龙, 王晗, 邹北骥. 基于注意力机制的医学影像深度哈希检索算法 Deep Hash Retrieval Algorithm for Medical Images Based on Attention Mechanism 计算机科学, 2022, 49(8): 113-119. https://doi.org/10.11896/jsjkx.210700153
[5]	檀莹莹, 王俊丽, 张超波. 基于图卷积神经网络的文本分类方法研究综述 Review of Text Classification Methods Based on Graph Convolutional Network 计算机科学, 2022, 49(8): 205-216. https://doi.org/10.11896/jsjkx.210800064
[6]	张颖涛, 张杰, 张睿, 张文强. 全局信息引导的真实图像风格迁移 Photorealistic Style Transfer Guided by Global Information 计算机科学, 2022, 49(7): 100-105. https://doi.org/10.11896/jsjkx.210600036
[7]	戴朝霞, 李锦欣, 张向东, 徐旭, 梅林, 张亮. 基于DNGAN的磁共振图像超分辨率重建算法 Super-resolution Reconstruction of MRI Based on DNGAN 计算机科学, 2022, 49(7): 113-119. https://doi.org/10.11896/jsjkx.210600105
[8]	刘月红, 牛少华, 神显豪. 基于卷积神经网络的虚拟现实视频帧内预测编码 Virtual Reality Video Intraframe Prediction Coding Based on Convolutional Neural Network 计算机科学, 2022, 49(7): 127-131. https://doi.org/10.11896/jsjkx.211100179
[9]	徐鸣珂, 张帆. Head Fusion:一种提高语音情绪识别的准确性和鲁棒性的方法 Head Fusion:A Method to Improve Accuracy and Robustness of Speech Emotion Recognition 计算机科学, 2022, 49(7): 132-141. https://doi.org/10.11896/jsjkx.210100085
[10]	金方焱, 王秀利. 融合RACNN和BiLSTM的金融领域事件隐式因果关系抽取 Implicit Causality Extraction of Financial Events Integrating RACNN and BiLSTM 计算机科学, 2022, 49(7): 179-186. https://doi.org/10.11896/jsjkx.210500190
[11]	张嘉淏, 刘峰, 齐佳音. 一种基于Bottleneck Transformer的轻量级微表情识别架构 Lightweight Micro-expression Recognition Architecture Based on Bottleneck Transformer 计算机科学, 2022, 49(6A): 370-377. https://doi.org/10.11896/jsjkx.210500023
[12]	王建明, 陈响育, 杨自忠, 史晨阳, 张宇航, 钱正坤. 不同数据增强方法对模型识别精度的影响 Influence of Different Data Augmentation Methods on Model Recognition Accuracy 计算机科学, 2022, 49(6A): 418-423. https://doi.org/10.11896/jsjkx.210700210
[13]	孙洁琪, 李亚峰, 张文博, 刘鹏辉. 基于离散小波变换的双域特征融合深度卷积神经网络 Dual-field Feature Fusion Deep Convolutional Neural Network Based on Discrete Wavelet Transformation 计算机科学, 2022, 49(6A): 434-440. https://doi.org/10.11896/jsjkx.210900199
[14]	杨玥, 冯涛, 梁虹, 杨扬. 融合交叉注意力机制的图像任意风格迁移 Image Arbitrary Style Transfer via Criss-cross Attention 计算机科学, 2022, 49(6A): 345-352. https://doi.org/10.11896/jsjkx.210700236
[15]	杨健楠, 张帆. 一种结合双注意力机制和层次网络结构的细碎农作物分类方法 Classification Method for Small Crops Combining Dual Attention Mechanisms and Hierarchical Network Structure 计算机科学, 2022, 49(6A): 353-357. https://doi.org/10.11896/jsjkx.210200169

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed