As of the last week of April, we received the credentials for the DH Canvas and we are now actively labeling the information we had on paper on the DH Canvas. We have also succeeded in extracting a lot of valuable information from the Venezia Città Mirabile that we will be adding on the DHCanvas. This article will describe how we are proceeding in entering the different landmarks on the DHCanvas.
Labeling in the DHCanvas
A high-resolution copy of the de’ Barbari map has been uploaded into the DHCanvas enabling us to annotate features on the map so that when a user searches a term, relevant images from the de’ Barbari map of Venice are also shown in the results. For example searching the term “Dogana da Mar” will return the image in Figure 1 along with any related annotated texts, enabling the user to get a sense of the spatial context during the 1500’s.
Figure 1: The annotation space illustrating the process for labelling Dogana da Mar in the DHCanvas.
Before entering a new name into DHCanvas we have to make sure that this name is not already present in the DHCanvas database, which includes trying all the possible variations of the name of a single place. That helps the DHCanvas search engine be as accurate as possible.
Since we were not able to find all the names of all the landmarks we entered on the DHCanvas, we made sure that at least every landmark has a “Tag” assigned to it. After browsing all the information we had, we settled upon 7 different tags to include in our labelings and those are :
- Bell Tower
The DHCanvas was designed for the annotation of text, which introduces some difficulties when labelling a large and complex image like the de’ Barbari map. Firstly the annotation tool only enables the annotation of square or rectangle features. Although this is adequate for capturing the rough location and size of a certain feature it can not be used to adequately identify specific features. An example of this is illustrated in Figure 2. This issue is significantly more problematic for certain features like canals. In the woodcut the canals are long and sinuous, curving all over the map and making it very difficult to label them adequately with a square. This problem is compounded by the fact that canals are likely heavily referenced in the other works in the DHCanvas- this annotated figure in the search results may be of little use to the user.
Figure 2: The annotation space for San Felice illustrating that the DHCanvas cannot accurately capture the true spatial extent of the island (given in red).
The DHCanvas is structured such that if multiple features are annotated one on top of another only the top layer can be accessed. This is not a serious problem, but it means that lower layers cannot be viewed or edited without first deleting the overlying annotations. Upon recognizing this we have been forced to modify our labelling process; we will first go through the entire map and annotate the largest features, generally the neighbourhoods, and then proceed to the next largest feature set.
Verification of labeling using the Venezia Città Mirabile and identifying additional features
The book Venezia Città Mirabile is a valuable source of information to verify the current labeling and to extract additional information. Indeed, it is organized by architectural features, like chimneys, islands, cemeteries, residential buildings… Some of the chapters are of great importance for our task: in particular the chapters on churches, convents, a canali, fondamente, calli, offer a listing of these features accompanied with pictures which are useful to determine their location on the map.
Once we have finished translating our annotation into the DHCanvas and checked our results over by comparing with the information from Venezia Città Mirabile, we will map our identified features onto a geographically referenced shapefile of Venice so that the annotations can be opened in Google maps and explored in a modern context.