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

Document images often have intricate layout structures, with numerous contentregions (e.g. texts, figures, tables) densely arranged on each page. This makesthe manual annotation of layout datasets expensive and inefficient. Thesecharacteristics also challenge existing active learning methods, as image-levelscoring and selection suffer from the overexposure of common objects.Inspiredby recent progresses in semi-supervised learning and self-training, we proposean Object-Level Active Learning framework for efficient document layoutAnnotation, OLALA. In this framework, only regions with the most ambiguousobject predictions within an image are selected for annotators to label,optimizing the use of the annotation budget. For unselected predictions, thesemi-automatic correction algorithm is proposed to identify certain errorsbased on prior knowledge of layout structures and rectifies them with minorsupervision. Additionally, we carefully design a perturbation-based objectscoring function for document images. It governs the object selection processvia evaluating prediction ambiguities, and considers both the positions andcategories of predicted layout objects. Extensive experiments show that OLALAcan significantly boost model performance and improve annotation efficiency,given the same labeling budget. Code for this paper can be accessed viahttps://github.com/lolipopshock/detectron2_al.

OLALA: Object-Level Active Learning for Efficient Document Layout Annotation