Structure extraction from PDF-based book documents

Abstract

Nowadays PDF documents have become a dominating knowledge repository for both the academia and industry largely because they are very convenient to print and exchange. However, the methods of automated structure information extraction are yet to be fully explored and the lack of effective methods hinders the information reuse of the PDF documents. To enhance the usability for PDF-formatted electronic books, we propose a novel computational framework to analyze the underlying physical structure and logical structure. The analysis is conducted at both page level and document level, including global typographies, reading order, logical elements, chapter/section hierarchy and metadata. Moreover, two characteristics of PDF-based books, i.e., style consistency in the whole book document and natural rendering order of PDF files, are fully exploited in this paper to improve the conventional image-based structure extraction methods. This paper employs the bipartite graph as a common structure for modeling various tasks, including reading order recovery, figure and caption association, and metadata extraction. Based on the graph representation, the optimal matching (OM) method is utilized to find the global optima in those tasks. Extensive benchmarking using real-world data validates the high efficiency and discrimination ability of the proposed method.

Extracting and matching authors and affiliations in scholarly documents

“…Such methods are brittle and may fail when presented with documents that follow new or different publishing styles. Gao et al [12] recently described SEB, a framework to detect the hierarchy and reading order of a document using weighted bipartite graphs but its fixed rules are again not flexible enough to capture document metadata in practical scenarios.…”

“…Systems by both Gao et al [12] and Lopez et al [21] correlate figures to their captions, enabling better document figure retrieval.…”

Transforming Japanese archives into accessible digital books

“…For example, Chaudhury et al [12] analyzed the layouts of scanned images of newspaper articles and automatically adds heading information. Also, Liangcai et al [13] proposed a system to automatically extract structures from a digital book format, PDF. These projects focus on automatically creating structures from scanned images or digital books.…”

Searching online book documents and analyzing book citations

“…Thus, it is now possible to find the references of a book during a book search. In addition, Thomson Reuters has released their Book Citation Index, giving researchers access to the citation network between books and the wider world of scholarly and scientific research and full bibliographies from books and book chapters 6 . However, to the best of our knowledge, there are no academic search engines or digital libraries that provide a citation list of a book that enables navigation to the sources cited in a book, even though this facility is typically available for papers (for example, in CiteSeer [9]).…”

“…It is still a challenging problem to design a ToC recognition algorithm that can be effectively applied to large scale heterogeneous books [14]. Gao et al studied both ToC and metadata extraction from PDF book documents by modeling them as a matching problem on the bipartite graph [6]. Feng et al studied how to restructure the OCR output of books using a Hidden Markov Model (HMM) based hierarchical alignment algorithm [5].…”