Rethinking Table Recognition using Graph Neural Networks

Qasim, Shah Rukh; Mahmood, Hassan

doi:10.1109/icdar.2019.00031

Cited by 119 publications

(82 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2) Graph Neural Networks So far in most of the mentioned approaches, the problem of segmenting tables in document images is treated with segmentation techniques. In 2019, Qasim et al [96] exploited the graph neural networks [97] to perform table recognition for the first time. The model is constructed with a blend of deep convolutional neural networks to extract image features and graph neural networks to control the relationship among the vertices.…”

Section: A: Fully Convolutional Networkmentioning

confidence: 99%

Current Status and Performance Analysis of Table Recognition in Document Images With Deep Neural Networks

et al. 2021

View full text Add to dashboard Cite

The first phase of table recognition is to detect the tabular area in a document. Subsequently, the tabular structures are recognized in the second phase in order to extract information from the respective cells. Table detection and structural recognition are pivotal problems in the domain of table understanding. However, table analysis is a perplexing task due to the colossal amount of diversity and asymmetry in tables. Therefore, it is an active area of research in document image analysis. Recent advances in the computing capabilities of graphical processing units have enabled the deep neural networks to outperform traditional state-of-the-art machine learning methods. Table understanding has substantially benefited from the recent breakthroughs in deep neural networks. However, there has not been a consolidated description of the deep learning methods for table detection and table structure recognition. This review paper provides a thorough analysis of the modern methodologies that utilize deep neural networks. This work provided a thorough understanding of the current state-of-the-art and related challenges of table understanding in document images. Furthermore, the leading datasets and their intricacies have been elaborated along with the quantitative results. Moreover, a brief overview is given regarding the promising directions that can serve as a guide to further improve table analysis in document images.

show abstract

Section: A: Fully Convolutional Networkmentioning

confidence: 99%

Current Status and Performance Analysis of Table Recognition in Document Images With Deep Neural Networks

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Recently, Chi et al [12] has exploited graph neural networks [13] to perform the task of table structure recognition on PDF documents. Another approach powered by graph neural networks is published by Qasim et al [38]. Their model combines the capabilities of convolutional neural networks and graph neural networks to extract tabular structures.…”

Section: B Deep Learning Based Approaches 1) Graph Neural Networkmentioning

confidence: 99%

Guided Table Structure Recognition Through Anchor Optimization

et al. 2021

View full text Add to dashboard Cite

This paper presents the novel approach towards table structure recognition by leveraging the guided anchors. The concept differs from current state-of-the-art systems for table structure recognition that naively apply object detection methods. In contrast to prior techniques, first, we estimate the viable anchors for table structure recognition. Subsequently, these anchors are exploited to locate the rows and columns in tabular images. Furthermore, the paper introduces a simple and effective method that improves the results using tabular layouts in realistic scenarios. The proposed method is exhaustively evaluated on the two publicly available datasets of table structure recognition: ICDAR-2013 and TabStructDB. Moreover, we empirically established the validity of our method by implementing it on the previous approaches. We accomplished state-of-the-art results on the ICDAR-2013 dataset with an average F-Measure of 94.19% (92.06% for rows and 96.32% for columns). Thus, a relative error reduction of more than 25% is achieved. Furthermore, our proposed post-processing improves the average F-Measure to 95.46% that results in a relative error reduction of more than 35%. Moreover, we surpassed the baseline results on the TabStructDB dataset with an average F-Measure of 94.57% (94.08% for rows and 95.06% for columns).

show abstract

“…Multi-modal methods have been proposed to extract predefined named entities from invoices [15,20]. Graph Neural Networks (GNNs) have been used to detect tables in invoice documents [24] as well as for association based table structure parsing [22]. Document classification is a partly related problem that has been studied using CNN-only approaches for identity document verification [26].…”

Section: Related Workmentioning

confidence: 99%

Multi-Modal Association based Grouping for Form Structure Extraction

Aggarwal

Sarkar

Gupta

et al. 2020

2020 IEEE Winter Conference on Applications of Computer Vision (WACV)

View full text Add to dashboard Cite

Document structure extraction has been a widely researched area for decades. Recent work in this direction has been deep learning-based, mostly focusing on extracting structure using fully convolution NN through semantic segmentation. In this work, we present a novel multi-modal approach for form structure extraction. Given simple elements such as textruns and widgets, we extract higher-order structures such as TextBlocks, Text Fields, Choice Fields, and Choice Groups, which are essential for information collection in forms. To achieve this, we obtain a local image patch around each low-level element (reference) by identifying candidate elements closest to it. We process textual and spatial representation of candidates sequentially through a BiLSTM to obtain context-aware representations and fuse them with image patch features obtained by processing it through a CNN. Subsequently, the sequential decoder takes this fused feature vector to predict the association type between reference and candidates. These predicted associations are utilized to determine larger structures through connected components analysis. Experimental results show the effectiveness of our approach achieving a recall of 90.29%, 73.80%, 83.12%, and 52.72% for the above structures, respectively, outperforming semantic segmentation baselines significantly. We show the efficacy of our method through ablations, comparing it against using individual modalities. We also introduce our new rich human-annotated Forms Dataset.

show abstract

Rethinking Table Recognition using Graph Neural Networks

Cited by 119 publications

References 32 publications

Current Status and Performance Analysis of Table Recognition in Document Images With Deep Neural Networks

Current Status and Performance Analysis of Table Recognition in Document Images With Deep Neural Networks

Guided Table Structure Recognition Through Anchor Optimization

Multi-Modal Association based Grouping for Form Structure Extraction

Contact Info

Product

Resources

About