English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Understanding documents with rich layouts is an essential step towardsinformation extraction. Business intelligence processes often require theextraction of useful semantic content from documents at a large scale forsubsequent decision-making tasks. In this context, instance-level segmentationof different document objects (title, sections, figures etc.) has emerged as aninteresting problem for the document analysis and understanding community. Toadvance the research in this direction, we present a transformer-based modelcalled \emph{DocSegTr} for end-to-end instance segmentation of complex layoutsin document images. The method adapts a twin attention module, for semanticreasoning, which helps to become highly computationally efficient compared withthe state-of-the-art. To the best of our knowledge, this is the first work ontransformer-based document segmentation. Extensive experimentation oncompetitive benchmarks like PubLayNet, PRIMA, Historical Japanese (HJ) andTableBank demonstrate that our model achieved comparable or better segmentationperformance than the existing state-of-the-art approaches with the averageprecision of 89.4, 40.3, 83.4 and 93.3. This simple and flexible frameworkcould serve as a promising baseline for instance-level recognition tasks indocument images.

DocSegTr: An Instance-Level End-to-End Document Image Segmentation Transformer