High-Definition Semantic Maps (Part III)
(continued from Part II)
This week I have toyed around with ways to zoom in and out of the map. After some experimentation I ended up with Seadragon, a Microsoft acquisitioned technology.
I have not used a very high resolution, but you can guess that one can delve rather deeply into the landscape. Which allows me to park there any documents, and even to show some preview.
Once one document reaches a size where one could read it, I thought that best is to start with the first page. Then, when the resolution allows, 2x2 pages are shown. At even more deeper levels 4x4 pages appear.
Certainly not in a quality which allows comfortable reading, but clear enough to get an idea and to check out the relevance of that document.
The documents themselves I generated from PDF as described earlier.
Application Specific Zoom
So this is not just one image thrown into the DeepZoom tile generation process. To make that happen, I implemented myself the DZI (DeepZoom Image) mechanism in Perl.
This also allows me to make the zoom process of the map more intelligent. What I can now do is to show more and more detail, the higher the zoom level is.