RE: [Bug-ocrad] Multi-Page PNM Files

[Kieran Hagzan]

Hi,

     I also must say -- hats off to you. I have been doing alot of research 
lately with OCR engines including many closed-source proprietary engines (as 
not to shame their developers, I will leave them nameless as I have found the 
general stability of the code unacceptable.) In this process I tend to do my 
own dithering, despeckling, and Minkowski dilation/erosion as the proprietary 
engines tend to have less memory issues (segfaults, leaks, etc. that way.) 
ocrad in most cases, produces text comparable with those expensive, proprietary 
OCR methods, and in some cases (namely color images that have been converted to 
bitonal) actually produces better results. I am impressed.

Excellent work,
-Kieran S. Hagzan


-----Original Message-----
From: Kieran Hagzan
Sent: Wed 3/8/2006 9:23 PM
To: Antonio Diaz Diaz; address@hidden
Subject: RE: [Bug-ocrad] Multi-Page PNM Files
 
Beautiful -- thank you.

-K


-----Original Message-----
From: Antonio Diaz Diaz [mailto:address@hidden
Sent: Wed 3/8/2006 8:28 PM
To: address@hidden
Cc: Kieran Hagzan
Subject: Re: [Bug-ocrad] Multi-Page PNM Files
 
Hello Kieran.

Kieran Hagzan wrote:
> Recent (development) versions of netpbm can take multi-page
> (multi-directory) TIFF images and create multi-page P*M files from
> them. Does ocrad by chance handle multipage documents?

Yes, ocrad 0.14 can read multipage PNM files. Either independet files 
concatenated together (every one with its own header), or merged under a 
unique header as pnmcat does.


Regards,
Antonio Diaz.