Automatic annotation with mapKurator - machines-reading-maps/Tutorials-Newsletters GitHub Wiki

This section of the tutorial explores the use of mapKurator to perform a large number of textual annotations on maps in a very short time. In its full version, mapKurator tries to match the words detected on maps to external URIs, such as Open Street Map nodes. However, in this platform, mapKurator is used simply to identify as many words as possible, and mark them as image annotations. Users can then verify and refine manually the annotations produced computationally, and save time from having to create each single annotation from scratch. In this instance, mapKurator is not using the user input to “learn” how to refine its word recognition process, but such functionalities are under development, and the manual annotations produced in this workshop will be used to improve mapKurator performance.