Addressing a Spatial Region

Use Case

You have a digitized issue of a newspaper. From that digitized object, there is a single article that you would like to display on its own in a viewer, rather than highlighting it with an Annotation in the context of the whole issue or even just in the context of its complete page.

Implementation Notes

You can focus on a region of a IIIF image or limit the display to a region in a few ways. This recipe focuses on one, but will describe the others briefly at the end. The way this recipe will address is that of pointing to the region of interest in a separate and independent manifest from the resource’s primary IIIF manifest. Approaching the use case this way leverages the abilities of IIIF to leave the original image intact, sharpens users’ focus on specific content, and by extension facilitates concentrated attention on annotations to the specified content.

To retrieve only a region of a IIIF resource, create an Annotation whose body contains the resource as a SpecificResource type. The source contains the resource’s details and a selector of type ImageApiSelector describes the region in a region property. The region value is the starting point and dimensions of the desired region, either as x,y,w,h (originating coordinates, width, and height) or as pct:x,y,w,h (originating coordinates, region width, and region height as percentages of the original resource).

Set the Canvas dimensions to equal the region you want to display, not the whole resource, to be sure the resource region displays in its original proportions. The region will be transformed as needed to fit the Canvas — enlarged or shrunk proportionally, or stretched or sqeezed disproportionally. (For more on the relationship of Canvas values and resource values, see Image and Canvas with Differing Dimensions.)

This recipe brings into relief the polyvocality of the term “Annotation”. On the one hand, the term and the IIIF construct can refer to presenting a IIIF resource or part of one in a viewer. This sort of Annotation, as in this recipe, takes the motivation value of painting and are conventionally called “painting Annotations”. The term can also be used to refer to associating additional data with all or part of a IIIF resource painted onto a Canvas. Because these latter kind of Annotations can have other values for their motivation property, they are frequently labeled “non-painting Annotations” as a group.

As mentioned above, there are other methods of drawing attention to a region of a IIIF resource:

  • Copy and Crop: Duplicate the image file you are working from, crop it in an image editor directly to the desired dimensions, and use this derived image’s URI in the Manifest. This method sidesteps the benefits of IIIF and walks right into one of the problems that IIIF aims to address, namely the proliferation of copies of high-quality images as image re-use grows.
  • Annotation: Create an Annotation on the full resource to tag the region. This method leaves the region in its resource context and allows you to draw attention to multiple regions without creating a new manifest for each.
  • Content State: Use the Content State API with a compatible viewer to show the region. This approach is usable only with a valid Content State URI and a IIIF viewer that works with Content State.
  • Static Image: Skip using an Image service and refer to your image resource file directly. This is as in Recipe #1, but here you would need to append a #xywh fragment identifier to the image URI. As with any IIIF use of a static image, the image will only be enlarged in the viewer, where an Image service offers among other benefits the download speed benefit of a tiled image, where the tiles are sized appropriate to zoom level and only delivered to the viewer when requested by the browser.


This recipe can not be applied to pulling a temporal segment from a time-based IIIF resource or to specifying a non-rectangular spatial region of an image. Incorporating a temporal region would use FragmentSelector and a non-rectangular region would use SvgSelector, but further description is beyond this recipe. However, see “Table of Contents for Multiple A/V Files on a Single Canvas” for an implementation example.


In this example we use an ImageApiSelector on the body of the Manifest’s sole Annotation to retrieve a single article selection from the 16 February 1925 issue of Berliner Tageblatt. The article discusses a meeting including Neville Chamberlain of Great Britain.

JSON-LD | View in Annona