(Archived) Problems parsing graphic text

[charles.roper 0]

Check out this screenshot:

http://www.flickr.com/photos/sxbrc_char ... 5/sizes/o/

Note how the word "salehurst" is found and highlighted. Now check this one:

http://www.flickr.com/photos/sxbrc_char ... 3/sizes/o/

Note how the text-based "MapMate" is being highlighted, but the graphic-based one is not. Generally I've found EN3 to be pretty good at recognising text, but it seems pretty erratic with this particular variety.