Eddy Bainbridge Posted October 14, 2017 Share Posted October 14, 2017 I have looked at some of the ineligible documents released, and feel there is sufficient data within them for restoration to be attempted using Optical Character Recognition Software. In the typewritten documents there are sufficient legible letters (or perhaps similar clearer documents could be used) to map the fonts and identify letters by means of elimination. The final weapon in the restoration armoury would be contextual evaluation (e.g if we get 'Ogwals' from the software, a human can have a good guess what the original word is). Has anyone tried this?, could it work? is there anyone in the research community who could try? Link to comment Share on other sites More sharing options...
Glenn Nall Posted October 15, 2017 Share Posted October 15, 2017 Eddy, are these documents images - jpgs or gifs or pngs? Are they pdfs? If you'd post a couple of examples that I can download I can try a couple of things. This kind of stuff always interests me... You can PM me and I'll give you my email address if you'd prefer that. I enjoy getting emails from real people who aren't trying to sell me Male Enhancement products or SootArounds. Link to comment Share on other sites More sharing options...
Eddy Bainbridge Posted October 15, 2017 Author Share Posted October 15, 2017 4 hours ago, Glenn Nall said: Eddy, are these documents images - jpgs or gifs or pngs? Are they pdfs? If you'd post a couple of examples that I can download I can try a couple of things. This kind of stuff always interests me... You can PM me and I'll give you my email address if you'd prefer that. I enjoy getting emails from real people who aren't trying to sell me Male Enhancement products or SootArounds. Hi Glenn, all docs are on the Mary Ferrell website. This looks like a straightforward to repair illegible document https://www.maryferrell.org/showDoc.html?docId=146988 This looks far harder https://www.maryferrell.org/showDoc.html?docId=147148 . Even the tough one, when you zoom in on the image appears to have dark pixels which MAY be able to be reformed into characters based on their estimated positions. I hope you have some success. Eddy Link to comment Share on other sites More sharing options...
Eddy Bainbridge Posted October 29, 2017 Author Share Posted October 29, 2017 (edited) I have tried to attach a CIA document https://www.maryferrell.org/showDoc.html?docId=146988 in Microsoft word format after my first attempt at repairing it. I decided my eyes were pretty good at character recognition but also that my brain might recognise patterns to fill in the blanks. As it is a CIA document I assumed it would contain cryptonyms but alas the only one I thought I could recognise was AMSPELL. I haven't looked deeply into the records released and some of the illegible documents don't look like they can be saved. I suspect some aren't even genuine text. The one I have chosen is tantalising to me, as I suspect someone with greater CIA document knowledge would see more in it than I do, and someone with greater I.T skills could take advantage of the visible marks that may be converted back into discernable letters. DOC 104-10211-10186.docx Edited October 29, 2017 by Eddy Bainbridge Link to comment Share on other sites More sharing options...
Glenn Nall Posted October 29, 2017 Share Posted October 29, 2017 3 hours ago, Eddy Bainbridge said: I have tried to attach a CIA document https://www.maryferrell.org/showDoc.html?docId=146988 in Microsoft word format after my first attempt at repairing it. I decided my eyes were pretty good at character recognition but also that my brain might recognise patterns to fill in the blanks. As it is a CIA document I assumed it would contain cryptonyms but alas the only one I thought I could recognise was AMSPELL. I haven't looked deeply into the records released and some of the illegible documents don't look like they can be saved. I suspect some aren't even genuine text. The one I have chosen is tantalising to me, as I suspect someone with greater CIA document knowledge would see more in it than I do, and someone with greater I.T skills could take advantage of the visible marks that may be converted back into discernable letters. DOC 104-10211-10186.docx see if these help (right-click, Save link as): cia-doc-contrasted-frosted-70.png cia-doc-contrasted-frosted-80.png cia-doc-contrasted.png Link to comment Share on other sites More sharing options...
Eddy Bainbridge Posted November 18, 2017 Author Share Posted November 18, 2017 With limited time this is the best I have managed. I guess to someone who knows what they are looking at it may spark an interest, or dismiss as a dudd. DOC 104-10211-10186.doc DOC 104-10211-10186.docx Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now