Table Structure Extraction with Bi-Directional Gated Recurrent Unit Networks

Abstract

Tables present summarized and structured information to the reader, which makes table's structure extraction an important part of document understanding applications. However, table structure identification is a hard problem not only because of the large variation in the table layouts and styles, but also owing to the variations in the page layouts and the noise contamination levels. A lot of research has been done to identify table structure, most of which is based on applying heuristics with the aid of optical character recognition (OCR) to hand pick layout features of the tables. These methods fail to generalize well because of the variations in the table layouts and the errors generated by OCR. In this paper, we have proposed a robust deep learning based approach to extract rows and columns from a detected table in document images with a high precision. In the proposed solution, the table images are first pre-processed and then fed to a bi-directional Recurrent Neural Network with Gated Recurrent Units (GRU) followed by a fully-connected layer with softmax activation. The network scans the images from top-to-bottom as well as left-to-right and classifies each input as either a row-separator or a column-separator. We have benchmarked our system on publicly available UNLV as well as ICDAR 2013 datasets on which it outperformed the state-of-theart table structure extraction systems by a significant margin.

What and Why? Towards Duo Explainable Fauxtography Detection Under Constrained Supervision

“…We build the TFE as an attention bi-directional GRU network [38], [39] to recurrently process word embeddings in each element of the text list and adaptively merges them to element-level features based on attention scores. In particular, we first construct a bi-direction GRU network to process word embedding sequences from both directions.…”

Guided Table Structure Recognition Through Anchor Optimization

“…Recurrent neural networks [40] have also been employed to handle the problem of table structure extraction [41], [42]. However, most of the prior approaches have utilized PDF meta-data.…”

Mixed Script Identification Using Automated DNN Hyperparameter Optimization

“…It can handle misspelled words and disambiguation problems also. LSTM [ 42 ], Bi-LSTM [ 43 ], GRU [ 44 ], and Bi-GRU [ 45 ] are investigated as classifiers (with one-dimensional convolution layer). These classifiers can process sequential data to overcome the short memory problem of recurrent neural networks.…”

“…A Long-Short Term Memory (LSTM) worked in one sequence or forward direction. According to [4,45,62], Bi-LSTM can capture or calculate both directions of contexts, such as upcoming and previous hidden layers. Backward layer:…”