Skip to main content
was successfully added to your cart.

Search, Export, and Labeling of Multi-page PDF Documents in the Annotation Lab

The Annotation supports labeling entities and relationships directly on multi-page PDF documents, and with release 2.5 has added support for search within PDF documents and export of annotated named entities in COCO format.

Search Through Visual NER Projects

For the Visual NER Projects, it is now possible to search for a keyword inside of image/pdf-based tasks using the search box available on the top of the Labeling page. Currently, the search is performed on the current page only.

Furthermore, we have also extended the keyword-based task search already available for text-based projects for Visual NER Projects.

On the Tasks page, use the search bar on the upper right side of the screen like you would do in other text-based projects, to identify all image/pdf tasks containing a given text.

visual_ner_search

Coco Export For Pdf Tasks In Visual NER Projects

Up until now, the COCO format export was limited to simple image documents. With the current version, this functionality is extended to single-page or multi-page pdf documents.

UI/UX Improvements

1. In Classification Project, the user is now able to use different layouts to show a list of choices.

dropdown_choices

  • layout=”select” : It will change choices from a list of choices inline to the dropdown layout. Possible values are “select”, “inline”, “vertical”
  • choice= “multiple”: Allow user to select multiple values from dropdown. Possible values are: “single”, “single-radio”, “multiple”

2. Better Toasts, Confirmation-Boxes, and Masking UI on potentially longer operations.

Below is the example, that shows toast and confirms box.

better-toast-confirm-box

Get & Install It HERE.

Full Feature Set HERE.

Rule-based annotations in the Annotation Lab

The Annotation Lab includes support for automated rule-based annotation - to enable programmatic labeling when training NLP models. You can seamlessly combine...