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

In recent years, tremendous efforts have been made on document imagerectification, but existing advanced algorithms are limited to processingrestricted document images, i.e., the input images must incorporate a completedocument. Once the captured image merely involves a local text region, itsrectification quality is degraded and unsatisfactory. Our previously proposedDocTr, a transformer-assisted network for document image rectification, alsosuffers from this limitation. In this work, we present DocTr++, a novel unifiedframework for document image rectification, without any restrictions on theinput distorted images. Our major technical improvements can be concluded inthree aspects. Firstly, we upgrade the original architecture by adopting ahierarchical encoder-decoder structure for multi-scale representationextraction and parsing. Secondly, we reformulate the pixel-wise mappingrelationship between the unrestricted distorted document images and thedistortion-free counterparts. The obtained data is used to train our DocTr++for unrestricted document image rectification. Thirdly, we contribute areal-world test set and metrics applicable for evaluating the rectificationquality. To our best knowledge, this is the first learning-based method for therectification of unrestricted document images. Extensive experiments areconducted, and the results demonstrate the effectiveness and superiority of ourmethod. We hope our DocTr++ will serve as a strong baseline for genericdocument image rectification, prompting the further advancement and applicationof learning-based algorithms. The source code and the proposed dataset arepublicly available at https://github.com/fh2019ustc/DocTr-Plus.

Deep Unrestricted Document Image Rectification