Processor: add dimension recognition

[sschober]

Currently, the image dimensions are hard coded; using pdfinfo, these could be inquired.

i fear this could add significant overhead to the translation process.

[sschober]

Two ideas:

1. add a different AST node type ImageWithMetaData which captures the dimensions
2. Add markdown syntax for meta data:[desc text]{320;200}(path) which directly produces the new type mentioned above