TabLeX: A Benchmark Dataset for Structure and Content Information Extraction from Scientific Tables
This addresses the challenge of information extraction from complex scientific tables for researchers and developers, but it is incremental as it builds on existing datasets and methods.
The paper introduces TabLeX, a benchmark dataset for extracting structure and content from scientific table images, highlighting that current state-of-the-art models fail even on simple tables.
Information Extraction (IE) from the tables present in scientific articles is challenging due to complicated tabular representations and complex embedded text. This paper presents TabLeX, a large-scale benchmark dataset comprising table images generated from scientific articles. TabLeX consists of two subsets, one for table structure extraction and the other for table content extraction. Each table image is accompanied by its corresponding LATEX source code. To facilitate the development of robust table IE tools, TabLeX contains images in different aspect ratios and in a variety of fonts. Our analysis sheds light on the shortcomings of current state-of-the-art table extraction models and shows that they fail on even simple table images. Towards the end, we experiment with a transformer-based existing baseline to report performance scores. In contrast to the static benchmarks, we plan to augment this dataset with more complex and diverse tables at regular intervals.