OCR Labeling
Last updated
Last updated
OCR (Optical Character Recognition) labeling is part of Span labeling, allowing you to label spans while viewing the original document as a reference. In OCR projects, you can also draw bounding boxes on the media file (e.g. PDF, TIF, JPG, PNG, GIF, DOCX, PPTX), link these bounding boxes to the transcription text, and view the text and the media file side-by-side.
OCR projects are helpful when working with image or document files that contain text. You can perform span labeling tasks like Named Entity Recognition (NER) and Part of Speech (POS) on the transcribed text. Additionally, you can annotate specific areas of the media files and link it to the corresponding text.
The layout of an OCR project is split into two sections: the media file displayed in the “Document Viewer” on the left, and the transcription text shown in the “Text Viewer” on the right.
You can shrink/enlarge one of the viewers by clicking and dragging the resize handler between the two viewers. You also can hide one of the viewers by clicking one of the arrows in the resize handler. This way, you can hide the document viewer if you wish to focus on labeling the text.
You’ll also notice a control panel at the top of the Document Viewer. This control panel has several buttons, and an indicator showing which page of the document you are viewing. The buttons on this control panel enable you to (from left to right): zoom in/out of the document, rotate the document counter-clockwise/clockwise, and draw bounding boxes.
Applying span labels in the Text Viewer will be familiar to users of a Span Labeling project. What makes an OCR project different is the capability to draw bounding boxes, and link those boxes to spans of text.
To draw a bounding box in the Document Viewer, click the rightmost icon on the Document Viewer’s control panel to enable drawing mode. The icon will turn blue when drawing mode is enabled.
In drawing mode, you can simply left-click and drag to draw the bounding boxes in the Document Viewer. Once you release the left-click, the bounding box will be drawn. A tooltip will then appear above the bounding box, telling you to select the text to bind it to the bounding box you just created.
When the bounding box is highlighted, you can select text in the Text Viewer. The span you select while the bounding box is highlighted will be bound to it. Whenever you click on the bounding box, the bound text will also be highlighted.
If you change your mind and want to bind the bounding box to a different text, simply right-click on the bounding box and select “Edit Sentence Position.” This will highlight the bounding box again, allowing you to select a different text to bind.
To adjust or resize the bounding box label size, simply left-click on the bounding box to display the resize handles. You can then drag the resize handles to adjust the size of the area.
Similar to span labels, bounding boxes can also cause conflicts in reviewer mode.
Conflicting bounding box labels have red lines and cannot be resized.
You can accept or reject them by right-clicking and choosing one of the options.
Bounding box labels that have reached the consensus have grey lines and can be resized.
Bounding box labels applied by reviewers have purple lines and can be resized.
Clicking on a conflict, consensus, or reviewer-applied bounding box label will highlight the connected bounding box. The same applies in reverse: clicking the bounding box will select the text and bounding box label it is bound to.
