[TUHS] tesseract has gotten so much better

[TUHS] tesseract has gotten so much better

[Will Senn]

[-- Attachment #1: Type: text/plain, Size: 1574 bytes --]

Hi All,

I just wanted to let y'all know that tesseract ocr has significantly 
improved and is much easier to use that it used to be. I have been using 
it with my workflow for a bit and it's crazy how much better it is than 
it was back when I  tried it last (admittedly 5-6 years ago). For those 
of you doing your own scans, or those of you finding sad little pdfs 
without ocr, the process is fairly simple.

Let's say you find "The Master Manual of Fortran.pdf" out there in the 
wild (or scan it). Here's how to turn it into a glorious ocr'd version:

Export your pdf as a multi-image tiff - it'll be ginormous, but you can 
delete it later (on Mac, this is just export from preview and select 
tiff, but gs will do it to, if I remember correctly) and then:

tesseract The\ Master\ Manual\ of\ Fortran.tiff out -l eng PDF

et voila, I nice, if large pdf, called out.pdf or somesuch will appear 
with ocr text that actually matches your scan (it seems to have caught 
up to adobe's ocr, or is quite close in my view, ymmv).

I speak English, so I installed tesseract and tesseract-eng, but it 
supports a bunch of other languages if you need them. Apparently 
google's been supporting and developing it for while now and if my 
results are any indicator, it's paying off (boy do I remember all the 
gobbledegook it used to produce).

tesseract will import from different image types, multiple images, etc. 
I just like the simplicity of tiff->pdf.

Anyhow, thought y'all might like to know as many of you live off the 
scans :).

Will

[-- Attachment #2: Type: text/html, Size: 2064 bytes --]