NSTL国家科技图书文献中心

1. VisFocus: Prompt-Guided Vision Encoders for OCR-Free Dense Document Understanding NSTL国家科技图书文献中心

Ofir Abramovich | Niv Nayman ... - 《Computer Vision - ECCV 2024,Part VIII》 - European Conference on Computer Vision - 2025, - 241～259 - 共19页

摘要： been made in the domain of visual document | understanding, with the prevailing architecture comprising a | the entire document. In this paper, we present | parts of the document, while disregarding others. We | document text fed to the visual encoder in place of the

关键词： Document understanding | OCR-free models

2. DLUE: Benchmarking Document Language Understanding NSTL国家科技图书文献中心

Ruoxi Xu | Hongyu Lin ... - 《Chinese Computational Linguistics》 - China National Conference on Computational Linguistics - 2025, - 387～401 - 共15页

摘要： to comprehensively evaluate document understanding | benchmark document* understanding researches, this paper* | structure and dispersed knowledge, document understanding | Understanding documents is central to many | summarizes four representative abilities, i.e., document

关键词： Document understanding | Language model | Evaluation

3. Perception-Enhanced Generative Transformer for Key Information Extraction from Documents NSTL国家科技图书文献中心

Runbo Zhao | Jun Jie Ou Yang ... - 《Pattern Recognition,Part XXXI》 - International Conference on Pattern Recognition - 2025, - 91～106 - 共16页

摘要： in document* images. To settle these issues, we* | Key information extraction (KIE) from scanned | documents has attracted significant attention due to | practical real-world applications. Despite impressive | results achieved by incorporating multimodal information

关键词： Key information extraction | Document understanding | Generative model

4. ROISER: Towards Real World Semantic Entity Recognition from Visually-Rich Documents NSTL国家科技图书文献中心

Zening Lin | Jiapeng Wang ... - 《Pattern Recognition,Part XXXI》 - International Conference on Pattern Recognition - 2025, - 76～90 - 共15页

摘要： the given visually-rich document* image, and it has* | Visual semantic entity recognition (visual SER | ) aims to extract contents that fall in key fields from | been widely applied across diverse scenarios. Most | existing visual SER methods employ the BIO tagging schema

关键词： Visual information extraction | Document understanding | Computer vision

5. Facet-Aware Multimodal Summarization via Cross-Modal Alignment NSTL国家科技图书文献中心

Yu Weng | Xuming Ye ... - 《Pattern Recognition,Part XIX》 - International Conference on Pattern Recognition - 2025, - 37～52 - 共16页

摘要： document segmentation module with a salient information | Multimodal generative models have demonstrated | promising capabilities for bridging the semantic gap | between visual and textual modalities, especially in the | context of multimodal summarization. Most of the

关键词： Multimedia analysis | Document understanding | Semantic technology | Summarization

6. LORE++: Logical location regression network for table structure recognition with pre-training NSTL国家科技图书文献中心

Long R. | Xing H.... - 《Pattern Recognition》 - 2025,157 - 共13页

摘要：© 2024 Elsevier LtdTable structure recognition | (TSR) aims at extracting tables in images into | machine-understandable formats. Current approaches | address this issue by either predicting the adjacency of | detected cells or direct generation of structural

关键词： OCR | Table structure recognition | Visual document understanding

7. C3E: A framework for chart classification and content extraction NSTL国家科技图书文献中心

Kanroo, Muhammad Suh... | Kawoosa, Hadia Showk...... - 《Computers and Electrical Engineering》 - 2025,121 - 共18页

摘要： challenge within the domain of document* analysis and* | understanding. The CCE problem can be viewed through a series | Incorporating charts into technical documents | enhances richness by simplifying complex data | representation and improving comprehension. However, automated

关键词： Document understanding | Object detection | Chart infographics | Computer vision | BVIP | Chart classification

8. Understanding document images by introducing explicit semantic information and short-range information interaction NSTL国家科技图书文献中心

Cheng, Yufeng | Wang, Dongxue ... - 《Image and vision computing》 - 2025,154(Feb.) - 1.1～1.14 - 共14页

摘要：Methods on the document* visual question* | , in this paper, we propose to utilize document | answering (DocVQA) task have achieved great success by | using pretrained multimodal models. However, two | issues are limiting their performances from further

关键词： DocVQA | Document semantic segmentation | Explicit semantic information | Star-shaped topology structure | Information interaction | MEANINGFUL USE

9. LiLTv2: Language-substitutable Layout-image Transformer for Visual Information Extraction NSTL国家科技图书文献中心

JIAPENG WANG | ZENING LIN ... - 《ACM transactions on multimedia computing communications and applications》 - 2025,21(3) - 72.1～72.27 - 共27页

摘要： to its pivotal role in intelligent document | Visual Information Extraction (VIE) has | experienced substantial growth and heightened interest due | processing. However, most existing related pre-trained | models typically can only process the data from a

关键词： Visual Information Extraction | Multi-Modal Document Understanding | Self-Supervised Pre-Training

10. Understanding the Lyαdocumentclass[12pt]{minimal} usepackage{amsmath} usepackage{wasysym} usepackage{amsfonts} usepackage{amssymb} usepackage{amsbsy} usepackage{mathrsfs} usepackage{upgreek} setlength{oddsidemargin}{-69pt} begin{document}$alpha $end{document} Emission Observed by the Solar Disk Imager Aboard the Advanced Space-Based Solar Observatory NSTL国家科技图书文献中心

Yiliang,Li | Ping,Zhang ... - 《Solar physics》 - 2025,300(6) - 共5页

摘要：{oddsidemargin}{-69pt} begin{document}$alpha $end{document | {document}$alpha $end{document} emission line across the | {oddsidemargin}{-69pt} begin{document}$alpha $end{document | {document}$alpha $end{document} line. Thus, the emission | {oddsidemargin}{-69pt} begin{document}$alpha $end{document

摘要： been made in the domain of visual document | understanding, with the prevailing architecture comprising a | the entire document. In this paper, we present | parts of the document, while disregarding others. We | document text fed to the visual encoder in place of the

关键词： Document understanding | OCR-free models

摘要： to comprehensively evaluate document understanding | benchmark document* understanding researches, this paper* | structure and dispersed knowledge, document understanding | Understanding documents is central to many | summarizes four representative abilities, i.e., document

关键词： Document understanding | Language model | Evaluation

摘要： in document* images. To settle these issues, we* | Key information extraction (KIE) from scanned | documents has attracted significant attention due to | practical real-world applications. Despite impressive | results achieved by incorporating multimodal information

关键词： Key information extraction | Document understanding | Generative model

摘要： the given visually-rich document* image, and it has* | Visual semantic entity recognition (visual SER | ) aims to extract contents that fall in key fields from | been widely applied across diverse scenarios. Most | existing visual SER methods employ the BIO tagging schema

关键词： Visual information extraction | Document understanding | Computer vision

摘要： document segmentation module with a salient information | Multimodal generative models have demonstrated | promising capabilities for bridging the semantic gap | between visual and textual modalities, especially in the | context of multimodal summarization. Most of the

关键词： Multimedia analysis | Document understanding | Semantic technology | Summarization

摘要：© 2024 Elsevier LtdTable structure recognition | (TSR) aims at extracting tables in images into | machine-understandable formats. Current approaches | address this issue by either predicting the adjacency of | detected cells or direct generation of structural

关键词： OCR | Table structure recognition | Visual document understanding

摘要： challenge within the domain of document* analysis and* | understanding. The CCE problem can be viewed through a series | Incorporating charts into technical documents | enhances richness by simplifying complex data | representation and improving comprehension. However, automated

关键词： Document understanding | Object detection | Chart infographics | Computer vision | BVIP | Chart classification

摘要：Methods on the document* visual question* | , in this paper, we propose to utilize document | answering (DocVQA) task have achieved great success by | using pretrained multimodal models. However, two | issues are limiting their performances from further

关键词： DocVQA | Document semantic segmentation | Explicit semantic information | Star-shaped topology structure | Information interaction | MEANINGFUL USE

摘要： to its pivotal role in intelligent document | Visual Information Extraction (VIE) has | experienced substantial growth and heightened interest due | processing. However, most existing related pre-trained | models typically can only process the data from a

关键词： Visual Information Extraction | Multi-Modal Document Understanding | Self-Supervised Pre-Training

关键词： Flares, spectrum | Spectral line, intensity and diagnostics | Prominences | Active regions, structure | Center-limb observations

4008-161-200 800-990-8900

国家科技图书文献中心

摘要： been made in the domain of visual document | understanding, with the prevailing architecture comprising a | the entire document. In this paper, we present | parts of the document, while disregarding others. We | document text fed to the visual encoder in place of the

关键词： Document understanding | OCR-free models

摘要： to comprehensively evaluate document understanding | benchmark document understanding researches, this paper | structure and dispersed knowledge, document understanding | Understanding documents is central to many | summarizes four representative abilities, i.e., document

关键词： Document understanding | Language model | Evaluation

摘要： in document images. To settle these issues, we | Key information extraction (KIE) from scanned | documents has attracted significant attention due to | practical real-world applications. Despite impressive | results achieved by incorporating multimodal information

关键词： Key information extraction | Document understanding | Generative model

摘要： the given visually-rich document image, and it has | Visual semantic entity recognition (visual SER | ) aims to extract contents that fall in key fields from | been widely applied across diverse scenarios. Most | existing visual SER methods employ the BIO tagging schema

关键词： Visual information extraction | Document understanding | Computer vision

摘要： document segmentation module with a salient information | Multimodal generative models have demonstrated | promising capabilities for bridging the semantic gap | between visual and textual modalities, especially in the | context of multimodal summarization. Most of the

关键词： Multimedia analysis | Document understanding | Semantic technology | Summarization

摘要：© 2024 Elsevier LtdTable structure recognition | (TSR) aims at extracting tables in images into | machine-understandable formats. Current approaches | address this issue by either predicting the adjacency of | detected cells or direct generation of structural

关键词： OCR | Table structure recognition | Visual document understanding

摘要： challenge within the domain of document analysis and | understanding. The CCE problem can be viewed through a series | Incorporating charts into technical documents | enhances richness by simplifying complex data | representation and improving comprehension. However, automated

关键词： Document understanding | Object detection | Chart infographics | Computer vision | BVIP | Chart classification

摘要：Methods on the document visual question | , in this paper, we propose to utilize document | answering (DocVQA) task have achieved great success by | using pretrained multimodal models. However, two | issues are limiting their performances from further

关键词： DocVQA | Document semantic segmentation | Explicit semantic information | Star-shaped topology structure | Information interaction | MEANINGFUL USE

摘要： to its pivotal role in intelligent document | Visual Information Extraction (VIE) has | experienced substantial growth and heightened interest due | processing. However, most existing related pre-trained | models typically can only process the data from a

关键词： Visual Information Extraction | Multi-Modal Document Understanding | Self-Supervised Pre-Training

关键词： Flares, spectrum | Spectral line, intensity and diagnostics | Prominences | Active regions, structure | Center-limb observations

4008-161-200 800-990-8900

国家科技图书文献中心

摘要： to comprehensively evaluate document understanding | benchmark document* understanding researches, this paper* | structure and dispersed knowledge, document understanding | Understanding documents is central to many | summarizes four representative abilities, i.e., document

摘要： in document* images. To settle these issues, we* | Key information extraction (KIE) from scanned | documents has attracted significant attention due to | practical real-world applications. Despite impressive | results achieved by incorporating multimodal information

摘要： the given visually-rich document* image, and it has* | Visual semantic entity recognition (visual SER | ) aims to extract contents that fall in key fields from | been widely applied across diverse scenarios. Most | existing visual SER methods employ the BIO tagging schema

摘要： challenge within the domain of document* analysis and* | understanding. The CCE problem can be viewed through a series | Incorporating charts into technical documents | enhances richness by simplifying complex data | representation and improving comprehension. However, automated

摘要：Methods on the document* visual question* | , in this paper, we propose to utilize document | answering (DocVQA) task have achieved great success by | using pretrained multimodal models. However, two | issues are limiting their performances from further