README.md

Table of contents

1. Papers
2. Datasets
3. Useful links

Papers

2022

• ERNIE-Layout: Layout Knowledge Enhanced Pre-training for Visually-rich Document Understanding,

  Qiming Peng, Yinxu Pan, Wenjin Wang, Bin Luo, Zhenyu Zhang, Zhengjie Huang, Teng Hu, Weichong Yin, Yongfeng Chen, Yin Zhang, Shikun Feng, Yu Sun, Hao Tian, Hua Wu, Haifeng Wang EMNLP (Findings) 2022

  Recent years have witnessed the rise and success of pre-training techniques in visually-rich document understanding. However, most existing methods lack the systematic mining and utilization of layout-centered knowledge, leading to sub-optimal performances. In this paper, we propose ERNIE-Layout, a novel document pre-training solution with layout knowledge enhancement in the whole workflow, to learn better representations that combine the features from text, layout, and image. Specifically, we first rearrange input sequences in the serialization stage, and then present a correlative pre-training task, reading order prediction, to learn the proper reading order of documents. To improve the layout awareness of the model, we integrate a spatial-aware disentangled attention into the multi-modal transformer and a replaced regions prediction task into the pre-training phase. Experimental results show that ERNIE-Layout achieves superior performance on various downstream tasks, setting new state-of-the-art on key information extraction, document image classification, and document question answering datasets.

• Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understanding,

  Kenton Lee, Mandar Joshi, Iulia Turc, Hexiang Hu, Fangyu Liu, Julian Eisenschlos, Urvashi Khandelwal, Peter Shaw, Ming-Wei Chang, Kristina Toutanova arxiv 2022

  Visually-situated language is ubiquitous -- sources range from textbooks with diagrams to web pages with images and tables, to mobile apps with buttons and forms. Perhaps due to this diversity, previous work has typically relied on domain-specific recipes with limited sharing of the underlying data, model architectures, and objectives. We present Pix2Struct, a pretrained image-to-text model for purely visual language understanding, which can be finetuned on tasks containing visually-situated language. Pix2Struct is pretrained by learning to parse masked screenshots of web pages into simplified HTML. The web, with its richness of visual elements cleanly reflected in the HTML structure, provides a large source of pretraining data well suited to the diversity of downstream tasks. Intuitively, this objective subsumes common pretraining signals such as OCR, language modeling, image captioning. In addition to the novel pretraining strategy, we introduce a variable-resolution input representation and a more flexible integration of language and vision inputs, where language prompts such as questions are rendered directly on top of the input image. For the first time, we show that a single pretrained model can achieve state-of-the-art results in six out of nine tasks across four domains: documents, illustrations, user interfaces, and natural images.

• XDoc: Unified Pre-training for Cross-Format Document Understanding,

  Jingye Chen, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei EMNLP 2022

  The surge of pre-training has witnessed the rapid development of document understanding recently. Pre-training and fine-tuning framework has been effectively used to tackle texts in various formats, including plain texts, document texts, and web texts. Despite achieving promising performance, existing pre-trained models usually target one specific document format at one time, making it difficult to combine knowledge from multiple document formats. To address this, we propose XDoc, a unified pre-trained model which deals with different document formats in a single model. For parameter efficiency, we share backbone parameters for different formats such as the word embedding layer and the Transformer layers. Meanwhile, we introduce adaptive layers with lightweight parameters to enhance the distinction across different formats. Experimental results have demonstrated that with only 36.7% parameters, XDoc achieves comparable or even better performance on a variety of downstream tasks compared with the individual pre-trained models, which is cost effective for real-world deployment.

• OCR-free Document Understanding Transformer,

  Geewook Kim, Teakgyu Hong, Moonbin Yim, Jeongyeon Nam, Jinyoung Park, Jinyeong Yim, Wonseok Hwang, Sangdoo Yun, Dongyoon Han, Seunghyun Park ECCV 2022

  Understanding document images (e.g., invoices) is a core but challenging task since it requires complex functions such as reading text and a holistic understanding of the document. Current Visual Document Understanding (VDU) methods outsource the task of reading text to off-the-shelf Optical Character Recognition (OCR) engines and focus on the understanding task with the OCR outputs. Although such OCR-based approaches have shown promising performance, they suffer from 1) high computational costs for using OCR; 2) inflexibility of OCR models on languages or types of document; 3) OCR error propagation to the subsequent process. To address these issues, in this paper, we introduce a novel OCR-free VDU model named Donut, which stands for Document understanding transformer. As the first step in OCR-free VDU research, we propose a simple architecture (i.e., Transformer) with a pre-training objective (i.e., cross-entropy loss). Donut is conceptually simple yet effective. Through extensive experiments and analyses, we show a simple OCR-free VDU model, Donut, achieves state-of-the-art performances on various VDU tasks in terms of both speed and accuracy. In addition, we offer a synthetic data generator that helps the model pre-training to be flexible in various languages and domains.

• LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking,

  Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei ACM Multimedia 2022

  Self-supervised pre-training techniques have achieved remarkable progress in Document AI. Most multimodal pre-trained models use a masked language modeling objective to learn bidirectional representations on the text modality, but they differ in pre-training objectives for the image modality. This discrepancy adds difficulty to multimodal representation learning. In this paper, we propose \textbf{LayoutLMv3} to pre-train multimodal Transformers for Document AI with unified text and image masking. Additionally, LayoutLMv3 is pre-trained with a word-patch alignment objective to learn cross-modal alignment by predicting whether the corresponding image patch of a text word is masked. The simple unified architecture and training objectives make LayoutLMv3 a general-purpose pre-trained model for both text-centric and image-centric Document AI tasks. Experimental results show that LayoutLMv3 achieves state-of-the-art performance not only in text-centric tasks, including form understanding, receipt understanding, and document visual question answering, but also in image-centric tasks such as document image classification and document layout analysis.

• LiLT: A Simple yet Effective Language-Independent Layout Transformer for Structured Document Understanding,

  Jiapeng Wang, Lianwen Jin, Kai Ding ACL 2022 Main conference 2022

  Structured document understanding has attracted considerable attention and made significant progress recently, owing to its crucial role in intelligent document processing. However, most existing related models can only deal with the document data of specific language(s) (typically English) included in the pre-training collection, which is extremely limited. To address this issue, we propose a simple yet effective Language-independent Layout Transformer (LiLT) for structured document understanding. LiLT can be pre-trained on the structured documents of a single language and then directly fine-tuned on other languages with the corresponding off-the-shelf monolingual/multilingual pre-trained textual models. Experimental results on eight languages have shown that LiLT can achieve competitive or even superior performance on diverse widely-used downstream benchmarks, which enables language-independent benefit from the pre-training of document layout structure. Code and model are publicly available at this https URL.

2021

• SelfDoc: Self-Supervised Document Representation Learning, [code\data ]

  Peizhao Li, Jiuxiang Gu, Jason Kuen, Vlad I. Morariu, Handong Zhao, Rajiv Jain, Varun Manjunatha, Hongfu Liu CVPR 2021

  We propose SelfDoc, a task-agnostic pre-training framework for document image understanding. Because documents are multimodal and are intended for sequential reading, our framework exploits the positional, textual, and visual information of every semantically meaningful component in a document, and it models the contextualization between each block of content. Unlike existing document pre-training models, our model is coarse-grained instead of treating individual words as input, therefore avoiding an overly fine-grained with excessive contextualization. Beyond that, we introduce cross-modal learning in the model pre-training phase to fully leverage multimodal information from unlabeled documents. For downstream usage, we propose a novel modality-adaptive attention mechanism for multimodal feature fusion by adaptively emphasizing language and vision signals. Our framework benefits from self-supervised pre-training on documents without requiring annotations by a feature masking training strategy. It achieves superior performance on multiple downstream tasks with significantly fewer document images used in the pre-training stage compared to previous works.

• LayoutReader: Pre-training of Text and Layout for Reading Order Detection, [code\data ]

  Zilong Wang, Yiheng Xu, Lei Cui, Jingbo Shang, Furu Wei EMNLP 2021

  Reading order detection is the cornerstone to understanding visually-rich documents (e.g., receipts and forms). Unfortunately, no existing work took advantage of advanced deep learning models because it is too laborious to annotate a large enough dataset. We observe that the reading order of WORD documents is embedded in their XML metadata; meanwhile, it is easy to convert WORD documents to PDFs or images. Therefore, in an automated manner, we construct ReadingBank, a benchmark dataset that contains reading order, text, and layout information for 500,000 document images covering a wide spectrum of document types. This first-ever large-scale dataset unleashes the power of deep neural networks for reading order detection. Specifically, our proposed LayoutReader captures the text and layout information for reading order prediction using the seq2seq model. It performs almost perfectly in reading order detection and significantly improves both open-source and commercial OCR engines in ordering text lines in their results in our experiments.

• MatchVIE: Exploiting Match Relevancy between Entities for Visual Information Extraction

  Guozhi Tang, Lele Xie, Lianwen Jin, Jiapeng Wang, Jingdong Chen, Zhen Xu, Qianying Wang, Yaqiang Wu, Hui Li IJCAI 2021

  Visual Information Extraction (VIE) task aims to extract key information from multifarious document images (e.g., invoices and purchase receipts). Most previous methods treat the VIE task simply as a sequence labeling problem or classification problem, which requires models to carefully identify each kind of semantics by introducing multimodal features, such as font, color, layout. But simply introducing multimodal features couldn't work well when faced with numeric semantic categories or some ambiguous texts. To address this issue, in this paper we propose a novel key-value matching model based on a graph neural network for VIE (MatchVIE). Through key-value matching based on relevancy evaluation, the proposed MatchVIE can bypass the recognitions to various semantics, and simply focuses on the strong relevancy between entities. Besides, we introduce a simple but effective operation, Num2Vec, to tackle the instability of encoded values, which helps model converge more smoothly. Comprehensive experiments demonstrate that the proposed MatchVIE can significantly outperform previous methods. Notably, to the best of our knowledge, MatchVIE may be the first attempt to tackle the VIE task by modeling the relevancy between keys and values and it is a good complement to the existing methods.

• StrucTexT: Structured Text Understanding with Multi-Modal Transformers

  Yulin Li, Yuxi Qian, Yuchen Yu, Xiameng Qin, Chengquan Zhang, Yan Liu, Kun Yao, Junyu Han, Jingtuo Liu, Errui Ding ACM Multimedia 2021

  Structured text understanding on Visually Rich Documents (VRDs) is a crucial part of Document Intelligence. Due to the complexity of content and layout in VRDs, structured text understanding has been a challenging task. Most existing studies decoupled this problem into two sub-tasks: entity labeling and entity linking, which require an entire understanding of the context of documents at both token and segment levels. However, little work has been concerned with the solutions that efficiently extract the structured data from different levels. This paper proposes a unified framework named StrucTexT, which is flexible and effective for handling both sub-tasks. Specifically, based on the transformer, we introduce a segment-token aligned encoder to deal with the entity labeling and entity linking tasks at different levels of granularity. Moreover, we design a novel pre-training strategy with three self-supervised tasks to learn a richer representation. StrucTexT uses the existing Masked Visual Language Modeling task and the new Sentence Length Prediction and Paired Boxes Direction tasks to incorporate the multi-modal information across text, image, and layout. We evaluate our method for structured text understanding at segment-level and token-level and show it outperforms the state-of-the-art counterparts with significantly superior performance on the FUNSD, SROIE, and EPHOIE datasets.

• DocFormer: End-to-End Transformer for Document Understanding

  Srikar Appalaraju, Bhavan Jasani, Bhargava Urala Kota, Yusheng Xie, R. Manmatha ICCV 2021

  We present DocFormer -- a multi-modal transformer based architecture for the task of Visual Document Understanding (VDU). VDU is a challenging problem which aims to understand documents in their varied formats (forms, receipts etc.) and layouts. In addition, DocFormer is pre-trained in an unsupervised fashion using carefully designed tasks which encourage multi-modal interaction. DocFormer uses text, vision and spatial features and combines them using a novel multi-modal self-attention layer. DocFormer also shares learned spatial embeddings across modalities which makes it easy for the model to correlate text to visual tokens and vice versa. DocFormer is evaluated on 4 different datasets each with strong baselines. DocFormer achieves state-of-the-art results on all of them, sometimes beating models 4x its size (in no. of parameters).

• Going Full-TILT Boogie on Document Understanding with Text-Image-Layout Transformer,

  Rafał Powalski, Łukasz Borchmann, Dawid Jurkiewicz, Tomasz Dwojak, Michał Pietruszka, Gabriela Pałka ICDAR 2021

  We address the challenging problem of Natural Language Comprehension beyond plain-text documents by introducing the TILT neural network architecture which simultaneously learns layout information, visual features, and textual semantics. Contrary to previous approaches, we rely on a decoder capable of unifying a variety of problems involving natural language. The layout is represented as an attention bias and complemented with contextualized visual information, while the core of our model is a pretrained encoder-decoder Transformer. Our novel approach achieves state-of-the-art results in extracting information from documents and answering questions which demand layout understanding (DocVQA, CORD, WikiOps, SROIE). At the same time, we simplify the process by employing an end-to-end model.

• LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding, [code ]

  Yang Xu et al. ACL 2021

  Pre-training of text and layout has proved effective in a variety of visually-rich document understanding tasks due to its effective model architecture and the advantage of large-scale unlabeled scanned/digital-born documents. In this paper, we present \textbf{LayoutLMv2} by pre-training text, layout and image in a multi-modal framework, where new model architectures and pre-training tasks are leveraged. Specifically, LayoutLMv2 not only uses the existing masked visual-language modeling task but also the new text-image alignment and text-image matching tasks in the pre-training stage, where cross-modality interaction is better learned. Meanwhile, it also integrates a spatial-aware self-attention mechanism into the Transformer architecture, so that the model can fully understand the relative positional relationship among different text blocks. Experiment results show that LayoutLMv2 outperforms strong baselines and achieves new state-of-the-art results on a wide variety of downstream visually-rich document understanding tasks, including FUNSD (0.7895 -> 0.8420), CORD (0.9493 -> 0.9601), SROIE (0.9524 -> 0.9781), Kleister-NDA (0.834 -> 0.852), RVL-CDIP (0.9443 -> 0.9564), and DocVQA (0.7295 -> 0.8672).

• LAMBERT: Layout-Aware (Language) Modeling using BERT for information extraction, [code ]

  Łukasz Garncarek, Rafał Powalski, Tomasz Stanisławek, Bartosz Topolski, Piotr Halama, Michał Turski, Filip Graliński ICDAR 2021

  In this paper we introduce a novel approach to the problem of understanding documents where the local semantics is influenced by non-trivial layout. Namely, we modify the Transformer architecture in a way that allows it to use the graphical features defined by the layout, without the need to re-learn the language semantics from scratch, thanks to starting the training process from a model pretrained on classical language modeling tasks. SOTA on [SROIE leaderboard](https://rrc.cvc.uab.es/?ch=13&com=evaluation&task=3)

• ViBERTgrid: A Jointly Trained Multi-Modal 2D Document Representation for Key Information Extraction from Documents

  Weihong Lin, Qifang Gao, Lei Sun, Zhuoyao Zhong, Kai Hu, Qin Ren, Qiang Huo ICDAR 2021

  Recent grid-based document representations like BERTgrid allow the simultaneous encoding of the textual and layout information of a document in a 2D feature map so that state-of-the-art image segmentation and/or object detection models can be straightforwardly leveraged to extract key information from documents. However, such methods have not achieved comparable performance to state-of-the-art sequence- and graph-based methods such as LayoutLM and PICK yet. In this paper, we propose a new multi-modal backbone network by concatenating a BERTgrid to an intermediate layer of a CNN model, where the input of CNN is a document image and the BERTgrid is a grid of word embeddings, to generate a more powerful grid-based document representation, named ViBERTgrid. Unlike BERTgrid, the parameters of BERT and CNN in our multimodal backbone network are trained jointly. Our experimental results demonstrate that this joint training strategy improves significantly the representation ability of ViBERTgrid. Consequently, our ViBERTgrid-based key information extraction approach has achieved state-of-the-art performance on real-world datasets.

• LAMPRET: Layout-Aware Multimodal PreTraining for Document Understanding

  Te-Lin Wu, Cheng Li, Mingyang Zhang, Tao Chen, Spurthi Amba Hombaiah, Michael Bendersky arxiv 2021

  Document layout comprises both structural and visual (eg. font-sizes) information that is vital but often ignored by machine learning models. The few existing models which do use layout information only consider textual contents, and overlook the existence of contents in other modalities such as images. Additionally, spatial interactions of presented contents in a layout were never really fully exploited. To bridge this gap, we parse a document into content blocks (eg. text, table, image) and propose a novel layout-aware multimodal hierarchical framework, LAMPreT, to model the blocks and the whole document. Our LAMPreT encodes each block with a multimodal transformer in the lower-level and aggregates the block-level representations and connections utilizing a specifically designed transformer at the higher-level. We design hierarchical pretraining objectives where the lower-level model is trained similarly to multimodal grounding models, and the higher-level model is trained with our proposed novel layout-aware objectives. We evaluate the proposed model on two layout-aware tasks -- text block filling and image suggestion and show the effectiveness of our proposed hierarchical architecture as well as pretraining techniques.

• LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding, [code/data ]

  Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei arxiv 2021

  Multimodal pre-training with text, layout, and image has achieved SOTA performance for visually-rich document understanding tasks recently, which demonstrates the great potential for joint learning across different modalities. In this paper, we present LayoutXLM, a multimodal pre-trained model for multilingual document understanding, which aims to bridge the language barriers for visually-rich document understanding. To accurately evaluate LayoutXLM, we also introduce a multilingual form understanding benchmark dataset named XFUN, which includes form understanding samples in 7 languages (Chinese, Japanese, Spanish, French, Italian, German, Portuguese), and key-value pairs are manually labeled for each language. Experiment results show that the LayoutXLM model has significantly outperformed the existing SOTA cross-lingual pre-trained models on the XFUN dataset.

• Glean: Structured Extractions from Templatic Documents

  Sandeep Tata et al. Proceedings of the VLDB Endowment 2021

  Extracting structured information from templatic documents is an important problem with the potential to automate many real-world business workflows such as payment, procurement, and payroll. The core challenge is that such documents can be laid out in virtually infinitely different ways. A good solution to this problem is one that generalizes well not only to known templates such as invoices from a known vendor, but also to unseen ones. We developed a system called Glean to tackle this problem. Given a target schema for a document type and some labeled documents of that type, Glean uses machine learning to automatically extract structured information from other documents of that type. In this paper, we describe the overall architecture of Glean, and discuss three key data management challenges : 1) managing the quality of ground truth data, 2) generating training data for the machine learning model using labeled documents, and 3) building tools that help a developer rapidly build and improve a model for a given document type. Through empirical studies on a real-world dataset, we show that these data management techniques allow us to train a model that is over 5 F1 points better than the exact same model architecture without the techniques we describe. We argue that for such information-extraction problems, designing abstractions that carefully manage the training data is at least as important as choosing a good model architecture.

• Improving Information Extraction from Visually Rich Documents using Visual Span Representations

  Ritesh Sarkhel, Arnab Nandi ResearchGate 2021

  Along with textual content, visual features play an essential role in the semantics of visually rich documents. Information extraction (IE) tasks perform poorly on these documents if these visual cues are not taken into account. In this paper, we present Artemis-a visually aware, machine-learning-based IE method for heterogeneous visually rich documents. Artemis represents a visual span in a document by jointly encoding its visual and textual context for IE tasks. Our main contribution is twofold. First, we develop a deep-learning model that identifies the local context boundary of a visual span with minimal human-labeling. Second, we describe a deep neural network that encodes the multimodal context of a visual span into a fixed-length vector by taking its textual and layout-specific features into account. It identifies the visual span(s) containing a named entity by leveraging this learned representation followed by an inference task. We evaluate Artemis on four heterogeneous datasets from different domains over a suite of information extraction tasks. Results show that it outperforms state-of-the-art text-based methods by up to 17 points in F1-score.

2020

• BROS: A Pre-trained Language Model for Understanding Texts in Document

  Teakgyu Hong, DongHyun Kim, Mingi Ji, Wonseok Hwang, Daehyun Nam, Sungrae Park openreview.net 2020

  Understanding document from their visual snapshots is an emerging and challenging problem that requires both advanced computer vision and NLP methods. Although the recent advance in OCR enables the accurate extraction of text segments, it is still challenging to extract key information from documents due to the diversity of layouts. To compensate for the difficulties, this paper introduces a pre-trained language model, BERT Relying On Spatiality (BROS), that represents and understands the semantics of spatially distributed texts. Different from previous pre-training methods on 1D text, BROS is pre-trained on large-scale semi-structured documents with a novel area-masking strategy while efficiently including the spatial layout information of input documents. Also, to generate structured outputs in various document understanding tasks, BROS utilizes a powerful graph-based decoder that can capture the relation between text segments. BROS achieves state-of-the-art results on four benchmark tasks: FUNSD, SROIE*, CORD, and SciTSR. Our experimental settings and implementation codes will be publicly available.

• LayoutLM: Pre-training of Text and Layout for Document Image Understanding, [code ]

  Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou KDD 2020

  Pre-training techniques have been verified successfully in a variety of NLP tasks in recent years. Despite the widespread of pre-training models for NLP applications, they almost focused on text-level manipulation, while neglecting the layout and style information that is vital for document image understanding. In this paper, we propose the LayoutLM to jointly model the interaction between text and layout information across scanned document images, which is beneficial for a great number of real-world document image understanding tasks such as information extraction from scanned documents. Furthermore, we also leverage the image features to incorporate the visual information of words into LayoutLM. To the best of our knowledge, this is the first time that text and layout are jointly learned in a single framework for document-level pre-training. It achieves new state-of-the-art results in several downstream tasks, including form understanding (from 70.72 to 79.27), receipt understanding (from 94.02 to 95.24) and document image classification (from 93.07 to 94.42). The code and pre-trained LayoutLM models are publicly available on GitHub.

• Representation Learning for Information Extraction from Form-like Documents, [code ]

  Bodhisattwa Prasad Majumder, Navneet Potti, Sandeep Tata, James Bradley Wendt, Qi Zhao, Marc Najork ACL 2020

  We propose a novel approach using representation learning for tackling the problem of extracting structured information from form-like document images. We propose an extraction system that uses knowledge of the types of the target fields to generate extraction candidates and a neural network architecture that learns a dense representation of each candidate based on neighboring words in the document. These learned representations are not only useful in solving the extraction task for unseen document templates from two different domains but are also interpretable, as we show using loss cases.

• PICK: Processing Key Information Extraction from Documents using Improved Graph Learning-Convolutional Networks, [code ]

  Wenwen Yu, Ning Lu, Xianbiao Qi, Ping Gong, Rong Xiao ICPR 2020

  Computer vision with state-of-the-art deep learning models has achieved huge success in the field of Optical Character Recognition (OCR) including text detection and recognition tasks recently. However, Key Information Extraction (KIE) from documents as the downstream task of OCR, having a large number of use scenarios in real-world, remains a challenge because documents not only have textual features extracting from OCR systems but also have semantic visual features that are not fully exploited and play a critical role in KIE. Too little work has been devoted to efficiently make full use of both textual and visual features of the documents. In this paper, we introduce PICK, a framework that is effective and robust in handling complex documents layout for KIE by combining graph learning with graph convolution operation, yielding a richer semantic representation containing the textual and visual features and global layout without ambiguity. Extensive experiments on real-world datasets have been conducted to show that our method outperforms baselines methods by significant margins.

• Attention-Based Graph Neural Network with Global Context Awareness for Document Understanding

  Yuan Hua, Z. Huang, J. Guo, Weidong Qiu CNCL 2020

  Information extraction from documents such as receipts or invoices is a fundamental and crucial step for office automation. Many approaches focus on extracting entities and relationships from plain texts, however, when it comes to document images, such demand becomes quite challenging since visual and layout information are also of great significance to help tackle this problem. In this work, we propose the attention-based graph neural network to combine textual and visual information from document images.Moreover, the global node is introduced in our graph construction algorithm which is used as a virtual hub to collect the information from all the nodes and edges to help improve the performance. Extensive experiments on real-world datasets show that our method outperforms baseline methods by significant margins.

• Towards a Multi-modal, Multi-task Learning based Pre-training Framework for Document Representation Learning

  Subhojeet Pramanik, Shashank Mujumdar, Hima Patel arxiv 2020

  In this paper, we propose a multi-task learning-based framework that utilizes a combination of self-supervised and supervised pre-training tasks to learn a generic document representation. We design the network architecture and the pre-training tasks to incorporate the multi-modal document information across text, layout, and image dimensions and allow the network to work with multi-page documents. We showcase the applicability of our pre-training framework on a variety of different real-world document tasks such as document classification, document information extraction, and document retrieval. We conduct exhaustive experiments to compare performance against different ablations of our framework and state-of-the-art baselines. We discuss the current limitations and next steps for our work.

• Merge and Recognize: A Geometry and 2D Context Aware Graph Model for Named Entity Recognition from Visual Documents

  Chuwei Luo, Yongpan Wang, Qi Zheng, Liangchen Li, Feiyu Gao, Shiyu Zhang COLING 2020

  Named entity recognition (NER) from visual documents, such as invoices, receipts or business cards, is a critical task for visual document understanding. Most classical approaches use a sequence-based model (typically BiLSTM-CRF framework) without considering document structure. Recent work on graph-based model using graph convolutional networks to encode visual and textual features have achieved promising performance on the task. However, few attempts take geometry information of text segments (text in bounding box) in visual documents into account. Meanwhile, existing methods do not consider that related text segments which need to be merged to form a complete entity in many real-world situations. In this paper, we present GraphNEMR, a graph-based model that uses graph convolutional networks to jointly merge text segments and recognize named entities. By incorporating geometry information from visual documents into our model, richer 2D context information is generated to improve document representations. To merge text segments, we introduce a novel mechanism that captures both geometry information as well as semantic information based on pre-trained language model. Experimental results show that the proposed GraphNEMR model outperforms both sequence-based and graph-based SOTA methods significantly.

• TRIE: End-to-End Text Reading and Information Extraction for Document Understanding

  Peng Zhang et al. Proceedings of the 28th ACM International Conference on Multimedia 2020

  Since real-world ubiquitous documents (e.g., invoices, tickets, resumes and leaflets) contain rich information, automatic document image understanding has become a hot topic. Most existing works decouple the problem into two separate tasks, (1) text reading for detecting and recognizing texts in images and (2) information extraction for analyzing and extracting key elements from previously extracted plain text.However, they mainly focus on improving information extraction task, while neglecting the fact that text reading and information extraction are mutually correlated. In this paper, we propose a unified end-to-end text reading and information extraction network, where the two tasks can reinforce each other. Specifically, the multimodal visual and textual features of text reading are fused for information extraction and in turn, the semantics in information extraction contribute to the optimization of text reading. On three real-world datasets with diverse document images (from fixed layout to variable layout, from structured text to semi-structured text), our proposed method significantly outperforms the state-of-the-art methods in both efficiency and accuracy.

• Robust Layout-aware IE for Visually Rich Documents with Pretrained Language Models

  Mengxi Wei, Yifan He, Qiong Zhang ACM SIGIR 2020

  Many business documents processed in modern NLP and IR pipelines are visually rich: in addition to text, their semantics can also be captured by visual traits such as layout, format, and fonts. We study the problem of information extraction from visually rich documents (VRDs) and present a model that combines the powerof large pre-trained language models and graph neural networks to efficiently encode both textual and visual information in business documents. We further introduce new fine-tuning objectives to improve in-domain unsupervised fine-tuning to better utilize large amount of unlabeled in-domain data. We experiment on real world invoice and resume data sets and show that the proposed method outperforms strong text-based RoBERTa baselines by 6.3% absolute F1 on invoices and 4.7% absolute F1 on resumes. When evaluated in a few-shot setting, our method requires up to 30x less annotation data than the baseline to achieve the same level of performance at∼90% F1.

• End-to-End Extraction of Structured Information from Business Documents with Pointer-Generator Networks

  Clément Sage et al. EMNLP 2020

  The predominant approaches for extracting key information from documents resort to classifiers predicting the information type of each word. However, the word level ground truth used for learning is expensive to obtain since it is not naturally produced by the extraction task. In this paper, we discuss a new method for training extraction models directly from the textual value of information. The extracted information of a document is represented as a sequence of tokens in the XML language. We learn to output this representation with a pointer-generator network that alternately copies the document words carrying information and generates the XML tags delimiting the types of information. The ability of our end-to-end method to retrieve structured information is assessed on a large set of business documents. We show that it performs competitively with a standard word classifier without requiring costly word level supervision.

• Information Extraction from Text Intensive and Visually Rich Banking Documents

  Berke Oral et al. Information Processing & Management 2020

  Document types, where visual and textual information plays an important role in their analysis and understanding, pose a new and attractive area for information extraction research. Although cheques, invoices, and receipts have been studied in some previous multi-modal studies, banking documents present an unexplored area due to the naturalness of the text they possess in addition to their visual richness. This article presents the first study which uses visual and textual information for deep-learning based information extraction on text-intensive and visually rich scanned documents which are, in this instance, unstructured banking documents, or more precisely, money transfer orders. The impact of using different neural word representations (i.e., FastText, ELMo, and BERT) on IE subtasks (namely, named entity recognition and relation extraction stages), positional features of words on document images and auxiliary learning with some other tasks are investigated. The article proposes a new relation extraction algorithm based on graph factorization to solve the complex relation extraction problem where the relations within documents are n-ary, nested, document-level, and previously indeterminate in quantity. Our experiments revealed that the use of deep learning algorithms yielded around 10 percentage points improvement on the IE sub-tasks. The inclusion of word positional features yielded around 3 percentage points of improvement in some specific information fields. Similarly, our auxiliary learning experiments yielded around 2 percentage points of improvement on some information fields associated with the specific transaction type detected by our auxiliary task. The integration of the information extraction system into a real banking environment reduced cycle times substantially. When compared to the manual workflow, document processing pipeline shortened book-to-book money transfers to 10 minutes (from 29 min.) and electronic fund transfers (EFT) to 17 minutes (from 41 min.) respectively.

2019

• Graph Convolution for Multimodal Information Extraction from Visually Rich Documents

  Xiaojing Liu, Feiyu Gao, Qiong Zhang, Huasha Zhao NAACL 2019

  Visually rich documents (VRDs) are ubiquitous in daily business and life. Examples are purchase receipts, insurance policy documents, custom declaration forms and so on. In VRDs, visual and layout information is critical for document understanding, and texts in such documents cannot be serialized into the one-dimensional sequence without losing information. Classic information extraction models such as BiLSTM-CRF typically operate on text sequences and do not incorporate visual features. In this paper, we introduce a graph convolution based model to combine textual and visual information presented in VRDs. Graph embeddings are trained to summarize the context of a text segment in the document, and further combined with text embeddings for entity extraction. Extensive experiments have been conducted to show that our method outperforms BiLSTM-CRF baselines by significant margins, on two real-world datasets. Additionally, ablation studies are also performed to evaluate the effectiveness of each component of our model.

• GraphIE: A Graph-Based Framework for Information Extraction, [code ]

  Yujie Qian, Enrico Santus, Zhijing Jin, Jiang Guo, Regina Barzilay NAACL 2019

  Most modern Information Extraction (IE) systems are implemented as sequential taggers and only model local dependencies. Non-local and non-sequential context is, however, a valuable source of information to improve predictions. In this paper, we introduce GraphIE, a framework that operates over a graph representing a broad set of dependencies between textual units (i.e. words or sentences). The algorithm propagates information between connected nodes through graph convolutions, generating a richer representation that can be exploited to improve word-level predictions. Evaluation on three different tasks — namely textual, social media and visual information extraction — shows that GraphIE consistently outperforms the state-of-the-art sequence tagging model by a significant margin.

• Attend, Copy, Parse: End-to-end information extraction from documents [code - unofficial ]

  Rasmus Berg Palm, Florian Laws, Ole Winther ICDAR 2019

  Document information extraction tasks performed by humans create data consisting of a PDF or document image input, and extracted string outputs. This end-to-end data is naturally consumed and produced when performing the task because it is valuable in and of itself. It is naturally available, at no additional cost. Unfortunately, state-of-the-art word classification methods for information extraction cannot use this data, instead requiring word-level labels which are expensive to create and consequently not available for many real life tasks. In this paper we propose the Attend, Copy, Parse architecture, a deep neural network model that can be trained directly on end-to-end data, bypassing the need for word-level labels. We evaluate the proposed architecture on a large diverse set of invoices, and outperform a state-of-the-art production system based on word classification. We believe our proposed architecture can be used on many real life information extraction tasks where word classification cannot be used due to a lack of the required word-level labels.

• One-shot Information Extraction from Document Images using Neuro-Deductive Program Synthesis

  Vishal Sunder, Ashwin Srinivasan, Lovekesh Vig, Gautam Shroff, Rohit Rahul arxiv 2019

  Our interest in this paper is in meeting a rapidly growing industrial demand for information extraction from images of documents such as invoices, bills, receipts etc. In practice users are able to provide a very small number of example images labeled with the information that needs to be extracted. We adopt a novel two-level neuro-deductive, approach where (a) we use pre-trained deep neural networks to populate a relational database with facts about each document-image; and (b) we use a form of deductive reasoning, related to meta-interpretive learning of transition systems to learn extraction programs: Given task-specific transitions defined using the entities and relations identified by the neural detectors and a small number of instances (usually 1, sometimes 2) of images and the desired outputs, a resource-bounded meta-interpreter constructs proofs for the instance(s) via logical deduction; a set of logic programs that extract each desired entity is easily synthesized from such proofs. In most cases a single training example together with a noisy-clone of itself suffices to learn a program-set that generalizes well on test documents, at which time the value of each entity is determined by a majority vote across its program-set. We demonstrate our two-level neuro-deductive approach on publicly available datasets ("Patent" and "Doctor's Bills") and also describe its use in a real-life industrial problem.

• EATEN: Entity-aware Attention for Single ShotVisual Text Extraction, [code ]

  He guo, Xiameng Qin, Jiaming Liu, Junyu Han, Jingtuo Liu, Errui Ding ICDAR 2019

  Extracting entity from images is a crucial part ofmany OCR applications, such as entity recognition of cards,invoices, and receipts. Most of the existing works employ classicaldetection and recognition paradigm. This paper proposes anEntity-aware Attention Text Extraction Network called EATEN,which is an end-to-end trainable system to extract the entitieswithout any post-processing. In the proposed framework, eachentity is parsed by its corresponding entity-aware decoder, re-spectively. Moreover, we innovatively introduce a state transitionmechanism which further improves the robustness of entityextraction. In consideration of the absence of public benchmarks,we construct a dataset of almost 0.6 million images in three real-world scenarios (train ticket, passport and business card), whichis publicly available at https://github.com/beacandler/EATEN. To the best of our knowledge, EATEN is the first single shotmethod to extract entities from images. Extensive experiments onthese benchmarks demonstrate the state-of-the-art performanceof EATEN.

• End-to-End Information Extraction byCharacter-Level Embedding and Multi-StageAttentional U-Net

  Tuan Nguyen Dang, Dat Nguyen Thanh BMVC 2019

  Information extraction from document images has received a lot of attention recently, due to the need for digitizing a large volume of unstructured documents such as invoices, receipts, bank transfers, etc. In this paper, we propose a novel deep learning architecture for end-to-end information extraction on the 2D character-grid embedding of the document, namely the Multi-Stage Attentional U-Net. To effectively capture the textual and spatial relations between 2D elements, our model leverages a specialized multi-stage encoder-decoders design, in conjunction with efficient uses of the self-attention mechanism and the box convolution. Experimental results on different datasets show that our model outperforms the baseline U-Net architecture by a large margin while using 40% fewer parameters. Moreover, it also significantly improved the baseline in erroneous OCR and limited training data scenario, thus becomes practical for real-world applications.

Older

• Chargrid: Towards Understanding 2D Documents, [code - unofficial ]

  Anoop R Katti et al. EMNLP 2018

  We introduce a novel type of text representation that preserves the 2D layout of a document. This is achieved by encoding each document page as a two-dimensional grid of characters. Based on this representation, we present a generic document understanding pipeline for structured documents. This pipeline makes use of a fully convolutional encoder-decoder network that predicts a segmentation mask and bounding boxes. We demonstrate its capabilities on an information extraction task from invoices and show that it significantly outperforms approaches based on sequential text or document images.

• CloudScan - A Configuration-Free Invoice Analysis System Using Recurrent Neural Networks, [code - unofficial ]

  Rasmus Berg Palm, Ole Winther, Florian Laws ICDAR 2017

  We present CloudScan; an invoice analysis system that requires zero configuration or upfront annotation. In contrast to previous work, CloudScan does not rely on templates of invoice layout, instead it learns a single global model of invoices that naturally generalizes to unseen invoice layouts. The model is trained using data automatically extracted from end-user provided feedback. This automatic training data extraction removes the requirement for users to annotate the data precisely. We describe a recurrent neural network model that can capture long range context and compare it to a baseline logistic regression model corresponding to the current CloudScan production system. We train and evaluate the system on 8 important fields using a dataset of 326,471 invoices. The recurrent neural network and baseline model achieve 0.891 and 0.887 average F1 scores respectively on seen invoice layouts. For the harder task of unseen invoice layouts, the recurrent neural network model outperforms the baseline with 0.840 average F1 compared to 0.788.

• Field Extraction by Hybrid Incremental and a-priori Structural Templates

  Vincent Poulain d'Andecy, Emmanuel Hartmann, Marçal Rusiñol DAS 2018

  In this paper, we present an incremental frame-work for extracting information fields from administrative documents. First, we demonstrate some limits of the existing state-of-the-art methods such as the delay of the system efficiency. This is a concern in industrial context when we have only few samples of each document class. Based on this analysis, we propose a hybrid system combining incremental learning by means of itf-df statistics and a-priori generic models. We report in the experimental section our results obtained with a dataset of real invoices.

• Multidomain Document Layout Understanding using Few Shot Object Detection

  Pranaydeep Singh, Srikrishna Varadarajan, Ankit Narayan Singh, Muktabh Mayank Srivastava arxiv 2018

  We try to address the problem of document layout understanding using a simple algorithm which generalizes across multiple domains while training on just few examples per domain. We approach this problem via supervised object detection method and propose a methodology to overcome the requirement of large datasets. We use the concept of transfer learning by pre-training our object detector on a simple artificial (source) dataset and fine-tuning it on a tiny domain specific (target) dataset. We show that this methodology works for multiple domains with training samples as less as 10 documents. We demonstrate the effect of each component of the methodology in the end result and show the superiority of this methodology over simple object detectors.

• Extracting structured data from invoices

  Xavier Holt, Andrew Chisholm ALTA 2018

  Business documents encode a wealth of information in a format tailored to human consumption – i.e. aesthetically disbursed natural language text, graphics and tables. We address the task of extracting key fields (e.g. the amount due on an invoice) from a wide-variety of potentially unseen document formats. In contrast to traditional template driven extraction systems, we introduce a content-driven machine-learning approach which is both robust to noise and generalises to unseen document formats. In a comparison of our approach with alternative invoice extraction systems, we observe an absolute accuracy gain of 20\% across compared fields, and a 25\%–94\% reduction in extraction latency.

• Automatic and interactive rule inference without ground truth

  Cérès Carton, Aurélie Lemaitre, Bertrand Coüasnon ICDAR 2015

  Dealing with non annotated documents for the design of a document recognition system is not an easy task. In general, statistical methods cannot learn without an annotated ground truth, unlike syntactical methods. However their ability to deal with non annotated data comes from the fact that the description is manually made by a user. The adaptation to a new kind of document is then tedious as the whole manual process of extraction of knowledge has to be redone. In this paper, we propose a method to extract knowledge and generate rules without any ground truth. Using large volume of non annotated documents, it is possible to study redundancies of some extracted elements in the document images. The redundancy is exploited through an automatic clustering algorithm. An interaction with the user brings semantic to the detected clusters. In this work, the extracted elements are some keywords extracted with word spotting. This approach has been applied to old marriage record field detection on the FamilySearch HIP2013 competition database. The results demonstrate that we successfully automatically infer rules from non annotated documents using the redundancy of extracted elements of the documents.

• Semantic Label and Structure Model based Approach for Entity Recognition in Database Context

  Nihel Kooli, Abdel Belaïd ICDAR 2015

  This paper proposes an entity recognition approach in scanned documents referring to their description in database records. First, using the database record values, the corresponding document fields are labeled. Second, entities are identified by their labels and ranked using a TF/IDF based score. For each entity, local labels are grouped into a graph. This graph is matched with a graph model (structure model) which represents geometric structures of local entity labels using a specific cost function. This model is trained on a set of well chosen entities semi-automatically annotated. At the end, a correction step allows us to complete the eventual entity mislabeling using geometrical relationships between labels. The evaluation on 200 business documents containing 500 entities reaches about 93% for recall and 97% for precision.

• Combining Visual and Textual Features for Information Extraction fromOnline Flyers

  Emilia Apostolova, Noriko Tomuro EMNLP 2014

  Information in visually rich formats such as PDF and HTML is often conveyed by a combination of textual and visual features. In particular, genres such as marketing flyers and info-graphics often augment textual information by its color, size, positioning, etc. As a result, traditional text-based approaches to information extraction (IE) could underperform. In this study, we present a supervised machine learning approach to IE from on-line commercial real estate flyers. We evaluated the performance of SVM classifiers on the task of identifying 12 types of named entities using a combination of textual and visual features. Results show that the addition of visual features such as color, size, and positioning significantly increased classifier performance.

• From one tree to a forest: a unified solution for structured web data extraction, [Website]

  Qiang Hao, Ruiqiong Chu Cai, Yanwei Pang, Lei Y Zhang ACM SIGIR 2011

  Structured data, in the form of entities and associated attributes, has been a rich web resource for search engines and knowledge databases. To efficiently extract structured data from enormous websites in various verticals (e.g., books, restaurants), much research effort has been attracted, but most existing approaches either require considerable human effort or rely on strong features that lack of flexibility. We consider an ambitious scenario -- can we build a system that (1) is general enough to handle any vertical without re-implementation and (2) requires only one labeled example site from each vertical for training to automatically deal with other sites in the same vertical? In this paper, we propose a unified solution to demonstrate the feasibility of this scenario. Specifically, we design a set of weak but general features to characterize vertical knowledge (including attribute-specific semantics and inter-attribute layout relationships). Such features can be adopted in various verticals without redesign; meanwhile, they are weak enough to avoid overfitting of the learnt knowledge to seed sites. Given a new unseen site, the learnt knowledge is first applied to identify page-level candidate attribute values, while inevitably involve false positives. To remove noise, site-level information of the new site is then exploited to boost up the true values. The site-level information is derived in an unsupervised manner, without harm to the applicability of the solution. Promising experimental performance on 80 websites in 8 distinct verticals demonstrated the feasibility and flexibility of the proposed solution.

• A probabilistic approach to printed document understanding

  Eric Medvet, Alberto Bartoli, Giorgio Davanzo IJDAR 2010

  We propose an approach for information extraction for multi-page printed document understanding. The approach is designed for scenarios in which the set of possible document classes, i.e., documents sharing similar content and layout, is large and may evolve over time. Describing a new class is a very simple task: the operator merely provides a few samples and then, by means of a GUI, clicks on the OCR-generated blocks of a document containing the information to be extracted. Our approach is based on probability: we derived a general form for the probability that a sequence of blocks contains the searched information. We estimate the parameters for a new class by applying the maximum likelihood method to the samples of the class. All these parameters depend only on block properties that can be extracted automatically from the operator actions on the GUI. Processing a document of a given class consists in finding the sequence of blocks, which maximizes the corresponding probability for that class. We evaluated experimentally our proposal using 807 multi-page printed documents of different domains (invoices, patents, data-sheets), obtaining very good results—e.g., a success rate often greater than 90% even for classes with just two samples.

Datasets

English

• DocILE Benchmark for Document Information Localization and Extraction, [Website] [benchmark] [code ]

  Štěpán Šimsa, Milan Šulc, Michal Uřičář, Yash Patel, Ahmed Hamdi, Matěj Kocián, Matyáš Skalický, Jiří Matas, Antoine Doucet, Mickaël Coustaty, Dimosthenis Karatzas arxiv pre-print 2023

  This paper introduces the DocILE benchmark with the largest dataset of business documents for the tasks of Key Information Localization and Extraction and Line Item Recognition. It contains 6.7k annotated business documents, 100k synthetically generated documents, and nearly~1M unlabeled documents for unsupervised pre-training. The dataset has been built with knowledge of domain- and task-specific aspects, resulting in the following key features: (i) annotations in 55 classes, which surpasses the granularity of previously published key information extraction datasets by a large margin; (ii) Line Item Recognition represents a highly practical information extraction task, where key information has to be assigned to items in a table; (iii) documents come from numerous layouts and the test set includes zero- and few-shot cases as well as layouts commonly seen in the training set. The benchmark comes with several baselines, including RoBERTa, LayoutLMv3 and DETR-based Table Transformer. These baseline models were applied to both tasks of the DocILE benchmark, with results shared in this paper, offering a quick starting point for future work. The dataset and baselines are available at this https URL.

• DUE: End-to-End Document Understanding Benchmark, [Website ]

  Łukasz Borchmann, Michał Pietruszka, Tomasz Stanislawek, Dawid Jurkiewicz, Michał Turski, Karolina Szyndler, Filip Graliński NIPS 2021

  Understanding documents with rich layouts plays a vital role in digitization and hyper-automation but remains a challenging topic in the NLP research community. Additionally, the lack of a commonly accepted benchmark made it difficult to quantify progress in the domain. To empower research in this field, we introduce the Document Understanding Evaluation (DUE) benchmark consisting of both available and reformulated datasets to measure the end-to-end capabilities of systems in real-world scenarios. The benchmark includes Visual Question Answering, Key Information Extraction, and Machine Reading Comprehension tasks over various document domains and layouts featuring tables, graphs, lists, and infographics. In addition, the current study reports systematic baselines and analyses challenges in currently available datasets using recent advances in layout-aware language modeling.

• Spatial Dual-Modality Graph Reasoning for Key Information Extraction, [Website ]

  Hongbin Sun, Zhanghui Kuang, Xiaoyu Yue, Chenhao Lin, Wayne Zhang arxiv 2021

  Key information extraction from document images is of paramount importance in office automation. Conventional template matching based approaches fail to generalize well to document images of unseen templates, and are not robust against text recognition errors. In this paper, we propose an end-to-end Spatial Dual-Modality Graph Reasoning method (SDMG-R) to extract key information from unstructured document images. We model document images as dual-modality graphs, nodes of which encode both the visual and textual features of detected text regions, and edges of which represent the spatial relations between neighboring text regions. The key information extraction is solved by iteratively propagating messages along graph edges and reasoning the categories of graph nodes. In order to roundly evaluate our proposed method as well as boost the future research, we release a new dataset named WildReceipt, which is collected and annotated tailored for the evaluation of key information extraction from document images of unseen templates in the wild. It contains 25 key information categories, a total of about 69000 text boxes, and is about 2 times larger than the existing public datasets. Extensive experiments validate that all information including visual features, textual features and spatial relations can benefit key information extraction. It has been shown that SDMG-R can effectively extract key information from document images of unseen templates, and obtain new state-of-the-art results on the recent popular benchmark SROIE and our WildReceipt. Our code and dataset will be publicly released.

• Kleister: Key Information Extraction Datasets Involving Long Documents with Complex Layouts, [data nda ], [data charity ]

  Tomasz Stanisławek, Filip Graliński, Anna Wróblewska, Dawid Lipiński, Agnieszka Kaliska, Paulina Rosalska, Bartosz Topolski, Przemysław Biecek ICDAR 2021

  Charity dataset size: Train(1 729), Dev(440), Test(609). NDA dataset size: Train(254), Dev(83), Test(203). Description: The relevance of Key Information Extraction (KIE) task is increasing in the natural language processing problems. But there are still only a few well-defined problems that serve as benchmarks for solutions in this area. To bridge this gap, we introduce two new datasets (Kleister NDA and Kleister Charity). They involve a mix of born-digital and scanned long formal documents in English. In these datasets, an NLP system is expected to find or infer various types of entities by utilizing both textual and structural layout features. The Kleister Charity dataset consists of 2,788 annual financial reports of charity organizations, i.e. 61,643 unique pages with 21,612 entities to extract. The Kleister NDA dataset contains 540 Non-disclosure Agreements, i.e. 3,229 unique pages with 2,160 entities to extract. We provide several state-of-the-art baseline systems from the KIE domain (Flair, BERT, RoBERTa, LayoutLM, LAMBERT), which show that our datasets pose a strong challenge to existing models. The best model achieved 81.77 % and 83.57 % F1-score on Kleister NDA and Kleister Charity datasets respectively. With this paper, we release our datasets to encourage progress on more in-depth and complex information extraction tasks.

• SROIE, [Website]

  Zheng Huang, Kai Chen, Jianhua He, Xiang Bai, Dimosthenis Karatzas, Shijian Lu, C. V. Jawahar ICDAR 2019

  Dataset size: Train(600), Test(400). Abstract: The dataset will have 1000 whole scanned receipt images. Each receipt image contains around about four key text fields, such as goods name, unit price and total cost, etc. The text annotated in the dataset mainly consists of digits and English characters. The dataset is split into a training/validation set (“trainval”) and a test set (“test”). The “trainval” set consists of 600 receipt images which will be made available to the participants along with their annotations. The “test” set consists of 400 images.

• CORD, [code/data ]

  Park, Seunghyun and Shin, Seung and Lee, Bado and Lee, Junyeop and Surh, Jaeheung and Seo, Minjoon and Lee, Hwalsuk NeurIPS Workshop Document Intelligence 2019

  Dataset size: Train(800), Dev(100), Test(100). Abstract: OCR is inevitably linked to NLP since its final output is in text. Advances in document intelligence are driving the need for a unified technology that integrates OCR with various NLP tasks, especially semantic parsing. Since OCR and semantic parsing have been studied as separate tasks so far, the datasets for each task on their own are rich, while those for the integrated post-OCR parsing tasks are relatively insufficient. In this study, we publish a consolidated dataset for receipt parsing as the first step towards post-OCR parsing tasks. The dataset consists of thousands of Indonesian receipts, which contains images and box/text annotations for OCR, and multi-level semantic labels for parsing. The proposed dataset can be used to address various OCR and parsing tasks.

• FUNSD, [Website]

  Guillaume Jaume, Hazım Kemal Ekenel, Jean-Philippe Thiran ICDAR-OST 2019

  Dataset size: Train(149), Test(50). Abstract: We present a new dataset for form understanding in noisy scanned documents (FUNSD) that aims at extracting and structuring the textual content of forms. The dataset comprises 199 real, fully annotated, scanned forms. The documents are noisy and vary widely in appearance, making form understanding (FoUn) a challenging task. The proposed dataset can be used for various tasks, including text detection, optical character recognition, spatial layout analysis, and entity labeling/linking. To the best of our knowledge, this is the first publicly available dataset with comprehensive annotations to address FoUn task. We also present a set of baselines and introduce metrics to evaluate performance on the FUNSD datase

• NIST, [Website]

  Darren Dimmick, Michael Garris, Charles Wilson, Patricia Flanagan

  The documents in this database are 12 different tax forms from the IRS 1040 Package X for the year 1988. These include Forms 1040, 2106, 2441, 4562, and 6251 together with Schedules A, B, C, D, E, F, and SE. There are 900 simulated tax submissions (Forms 1040, 2106, 2441, 4562, and 6251 together with Schedules A, B, C, D, E, F, and SE). Suitable for both document processing and automated data capture research, development, and evaluation, the data set can be used for: a) forms identification, b) field isolation; locating the entry fields on the form, c) character segmentation: separating entry field values into characters, d) character recognition: identifying specific machine printed characters

• Deepform, [Website]

  Jonathan Stray, Nicholas Bardy

  DeepForm aims to extract information from TV and cable political advertising disclosure forms using deep learning and provide a challenging journalism-relevant dataset for NLP/ML researchers. This public data is valuable to journalists but locked in PDFs. Through this benchmark, we hope to accelerate collaboration on the concrete task of making this data accessible and longer-term solutions for general information extraction from visually-structured documents in fields like medicine, climate science, social science, and beyond.

Chinese

• (EPHOIE) Towards Robust Visual Information Extraction in Real World: New Dataset and Novel Solution, [code/data ]

  Jiapeng Wang, Chongyu Liu, Lianwen Jin, Guozhi Tang, Jiaxin Zhang, Shuaitao Zhang, Qianying Wang, Yaqiang Wu, Mingxiang Cai AAAI 2021

  Visual information extraction (VIE) has attracted considerable attention recently owing to its various advanced applications such as document understanding, automatic marking and intelligent education. Most existing works decoupled this problem into several independent sub-tasks of text spotting (text detection and recognition) and information extraction, which completely ignored the high correlation among them during optimization. In this paper, we propose a robust visual information extraction system (VIES) towards real-world scenarios, which is a unified end-to-end trainable framework for simultaneous text detection, recognition and information extraction by taking a single document image as input and outputting the structured information. Specifically, the information extraction branch collects abundant visual and semantic representations from text spotting for multimodal feature fusion and conversely, provides higher-level semantic clues to contribute to the optimization of text spotting. Moreover, regarding the shortage of public benchmarks, we construct a fully-annotated dataset called EPHOIE (this https URL), which is the first Chinese benchmark for both text spotting and visual information extraction. EPHOIE consists of 1,494 images of examination paper head with complex layouts and background, including a total of 15,771 Chinese handwritten or printed text instances. Compared with the state-of-the-art methods, our VIES shows significant superior performance on the EPHOIE dataset and achieves a 9.01% F-score gain on the widely used SROIE dataset under the end-to-end scenario.

• Metaknowledge Extraction Based onMulti-Modal Documents, [code/data ]

  Shukan Liu, Ruilin Xu, Boying Geng, Qiao Sun, Li Duan, Yiming Liu IEEE Access 2011

  The triplet-based knowledge in large-scale knowledge bases is most likely lacking in structural logic and problematic of conducting knowledge hierarchy. In this paper, we introduce the concept of metaknowledge to knowledge engineering research for the purpose of structural knowledge construction. Therefore, the Metaknowledge Extraction Framework and Document Structure Tree model are presented to extract and organize metaknowledge elements (titles, authors, abstracts, sections, paragraphs, etc.), so that it is feasible to extract the structural knowledge from multi-modal documents. Experiment results have proved the effectiveness of metaknowledge elements extraction by our framework. Meanwhile, detailed examples are given to demonstrate what exactly metaknowledge is and how to generate it. At the end of this paper, we propose and analyze the task flow of metaknowledge applications and the associations between knowledge and metaknowledge.

• EATEN: Entity-aware Attention for Single Shot Visual Text Extraction, [data], [code ]

  He Guo, Xiameng Qin, Jiaming Liu, Junyu Han, Jingtuo Liu and Errui Ding ICDAR 2019

  Abstract: Extracting entity from images is a crucial part of many OCR applications, such as entity recognition of cards, invoices, and receipts. Most of the existing works employ classical detection and recognition paradigm. This paper proposes an Entity-aware Attention Text Extraction Network called EATEN, which is an end-to-end trainable system to extract the entities without any post-processing. In the proposed framework, each entity is parsed by its corresponding entity-aware decoder, respectively. Moreover, we innovatively introduce a state transition mechanism which further improves the robustness of entity extraction. In consideration of the absence of public benchmarks, we construct a dataset of almost 0.6 million images in three real-world scenarios (train ticket, passport and business card), which is publicly available at this https URL. To the best of our knowledge, EATEN is the first single shot method to extract entities from images. Extensive experiments on these benchmarks demonstrate the state-of-the-art performance of EATEN.

Polish

• Results of the PolEval 2020 Shared Task 4: Information Extraction from Long Documents with Complex Layouts, [Website]
  Filip Graliński, Anna Wróblewska Proceedings of the PolEval 2020 Workshop 2020

  Dataset size: Train(1 628), Dev(548), Test(555). Description: The challenge is about information acquisition and inference in the field of natural language processing. Collecting information from real, long documents must deal with complex page layouts by integrating found entities along multiple pages and text sections, tables, plots, forms, etc. To encourage progress in deeper and more complex information extraction, we present a dataset in which systems have to find the most important information about different types of entities from formal documents. These units are not only classes from the systems for recognising units with a standard name (NER) (e.g. person, location or organisation), but also the roles of units in whole documents (e.g. chairman of the board, date of issue).

Multilanguage

• XFUND: A Benchmark Dataset for Multilingual Visually Rich Form Understanding

  Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei ACL 2022

  Multimodal pre-training with text, layout, and image has achieved SOTA performance for visually rich document understanding tasks recently, which demonstrates the great potential for joint learning across different modalities. However, the existed research work has focused only on the English domain while neglecting the importance of multilingual generalization. In this paper, we introduce a human-annotated multilingual form understanding benchmark dataset named XFUND, which includes form understanding samples in 7 languages (Chinese, Japanese, Spanish, French, Italian, German, Portuguese). Meanwhile, we present LayoutXLM, a multimodal pre-trained model for multilingual document understanding, which aims to bridge the language barriers for visually rich document understanding. Experimental results show that the LayoutXLM model has significantly outperformed the existing SOTA cross-lingual pre-trained models on the XFUND dataset.

• Ghega dataset, [Website]

  Vishal Sunder, Ashwin Srinivasan, Lovekesh Vig, Gautam Shroff, Rohit Rahul arxiv 2019

  The dataset is composed as follows. It contains two groups of documents: 110 data-sheets of electronic components and 136 patents. Each group is further divided in classes: data-sheets classes share the component type and producer; patents classes share the patent source.

Useful links

Data Augmentation:

1. https://github.com/makcedward/nlpaug
2. https://github.com/dsfsi/textaugment
3. https://github.com/QData/TextAttack
4. https://github.com/joke2k/faker
5. https://github.com/benkeen/generatedata
6. https://github.com/Belval/TextRecognitionDataGenerator
7. https://github.com/snorkel-team/snorkel

Related NLP topics:

1. Named Entity Recognition (NER)
2. Entity Linking (EL)
3. Template extraction
4. Noun Phrase Canonicalization

Others:

1. cleanlab - The standard package for machine learning with noisy labels and finding mislabeled data. Works with most datasets and models.
2. CommonRegex - find all times, dates, links, phone numbers, emails, ip addresses, prices, hex colors, and credit card numbers in a string
3. Name Parser - parsing human names into their individual components
4. pyahocorasick - is a fast and memory efficient library for exact or approximate multi-pattern string search meaning that you can find multiple key strings occurrences at once in some input text
5. deepmatcher - performing entity and text matching using deep learning
6. simstring - A Python implementation of the SimString, a simple and efficient algorithm for approximate string matching.
7. RapidFuzz - RapidFuzz is a fast string matching library for Python and C++, which is using the string similarity calculations from FuzzyWuzzy
8. bootleg - self-supervised named entity disambiguation (NED) system that links mentions in text to entities in a knowledge base