Deep Learning in Time-Frequency Domain for Document Layout Analysis

Document layout analysis plays an important role in the area of Document Understanding. It is responsible for identifying and classifying the different components of digital documents. Currently, there is no universal algorithm that fits all types of digital documents. This work presents a novel app...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Felipe Grijalva, Erick Santos, Byron Acuna, Juan Carlos Rodriguez, Julio Cesar Larco
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Computer-vision deep-learning document layout analysis feature engineering PDF Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/9d207e9589d64e068014822efdf7aca1
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:9d207e9589d64e068014822efdf7aca1
record_format	dspace
spelling	oai:doaj.org-article:9d207e9589d64e068014822efdf7aca12021-11-17T00:01:12ZDeep Learning in Time-Frequency Domain for Document Layout Analysis2169-353610.1109/ACCESS.2021.3125913https://doaj.org/article/9d207e9589d64e068014822efdf7aca12021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9605682/https://doaj.org/toc/2169-3536Document layout analysis plays an important role in the area of Document Understanding. It is responsible for identifying and classifying the different components of digital documents. Currently, there is no universal algorithm that fits all types of digital documents. This work presents a novel approach for identifying tables, figures, isolated equations and text regions in scientific papers using deep learning and computer vision techniques. Our proposed approach is a three-stage system: (i) obtaining the spectrograms of the horizontal and vertical intensity histograms of segmented regions of interest; (ii) labeling segmented regions of interest into text, table, and figures using a deep convolutional neural network classifier; and (iii) identifying isolated equations in text regions using Bag of Visual Words (BOVW) with Zernike moments. We built a new dataset composed of 11007 papers to perform the experiments, using two common segmentation metrics to evaluate our model: (1) Adjusted Rand Index (ARI) and (2) Variation of Information (VI). The proposed document layout analysis system reached an overall accuracy of 96.2685%, outperforming prior art with a less computational cost.Felipe GrijalvaErick SantosByron AcunaJuan Carlos RodriguezJulio Cesar LarcoIEEEarticleComputer-visiondeep-learningdocument layout analysisfeature engineeringPDFElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 151254-151265 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Computer-vision deep-learning document layout analysis feature engineering PDF Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Computer-vision deep-learning document layout analysis feature engineering PDF Electrical engineering. Electronics. Nuclear engineering TK1-9971 Felipe Grijalva Erick Santos Byron Acuna Juan Carlos Rodriguez Julio Cesar Larco Deep Learning in Time-Frequency Domain for Document Layout Analysis
description	Document layout analysis plays an important role in the area of Document Understanding. It is responsible for identifying and classifying the different components of digital documents. Currently, there is no universal algorithm that fits all types of digital documents. This work presents a novel approach for identifying tables, figures, isolated equations and text regions in scientific papers using deep learning and computer vision techniques. Our proposed approach is a three-stage system: (i) obtaining the spectrograms of the horizontal and vertical intensity histograms of segmented regions of interest; (ii) labeling segmented regions of interest into text, table, and figures using a deep convolutional neural network classifier; and (iii) identifying isolated equations in text regions using Bag of Visual Words (BOVW) with Zernike moments. We built a new dataset composed of 11007 papers to perform the experiments, using two common segmentation metrics to evaluate our model: (1) Adjusted Rand Index (ARI) and (2) Variation of Information (VI). The proposed document layout analysis system reached an overall accuracy of 96.2685%, outperforming prior art with a less computational cost.
format	article
author	Felipe Grijalva Erick Santos Byron Acuna Juan Carlos Rodriguez Julio Cesar Larco
author_facet	Felipe Grijalva Erick Santos Byron Acuna Juan Carlos Rodriguez Julio Cesar Larco
author_sort	Felipe Grijalva
title	Deep Learning in Time-Frequency Domain for Document Layout Analysis
title_short	Deep Learning in Time-Frequency Domain for Document Layout Analysis
title_full	Deep Learning in Time-Frequency Domain for Document Layout Analysis
title_fullStr	Deep Learning in Time-Frequency Domain for Document Layout Analysis
title_full_unstemmed	Deep Learning in Time-Frequency Domain for Document Layout Analysis
title_sort	deep learning in time-frequency domain for document layout analysis
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/9d207e9589d64e068014822efdf7aca1
work_keys_str_mv	AT felipegrijalva deeplearningintimefrequencydomainfordocumentlayoutanalysis AT ericksantos deeplearningintimefrequencydomainfordocumentlayoutanalysis AT byronacuna deeplearningintimefrequencydomainfordocumentlayoutanalysis AT juancarlosrodriguez deeplearningintimefrequencydomainfordocumentlayoutanalysis AT juliocesarlarco deeplearningintimefrequencydomainfordocumentlayoutanalysis
_version_	1718426065736564736

Deep Learning in Time-Frequency Domain for Document Layout Analysis

Ejemplares similares