Bulk background filling on 300+ pages

[muguk]

I'm currently wading through individual pages of a scanned manual (an old Atari ST one for Hisoft 'C') which is 380+ pages long. Adobe Writer tried it's best to OCR it on the first pass but certain words kept coming up as unrecognised due to the bad scan/yellowed hue in the background. When I remove the yellow background, the OCR operation works a lot better.

I'm using Paint.NET's fill option on the background (white) with 50% tolerance and it's working a treat but is there any way to automate this process? There are a few other manuals which could do with some prep work done on them too before I attempt to OCR those as well and they have the same yellowing background.

Also, for the few pages that are a bad scan (I didn't do the original), how can I 'liven' up the text? My first attempt was using the colour replace function to replace the various shades of grey in the scan with a much darker one. Is there a plugin / option I can use to only sharpen up the text?

[yellowman]

Try Thrshold plugin, find it in this package :

http://paintdotnet.forumer.com/viewtopic.php?f=16&t=26244

[Rick Brewster]

The current version of Paint.NET can't be used for batch processing. Sorry.

[Ego Eram Reputo]

However Scriptlab can be used to automate a number of steps if you're applying the same methodology to each image. It's not batch processing, but it might help.

I'd also have a look at the Alphaspace plugin for background removal: viewtopic.php?f=16&t=27959&start=0

[muguk]

In theory, the top 10x10 pixels of each scanned image will be exactly the same - however <Family-friendly forum.>-eyed the original scan was.

Will take a look at this tomorrow in work as that's where I've left all the project files.