=>

+

This is a tool for processing scanned documents in tiff format. It works with multipage tiffs and offers two functions:

• extracting two text pages (or one text page) from one image as shown above,
• noise filtering so that the documents get smaller

This is a hack (I have modified tiffcp) with limited functioning and no working guarantee at all! In particular, it requires CCITT-compressed b/w images (the standard way how monochrome images such as scans are stored). It works with Windows (precompiled available) and Unix/Linux. The interface is only command line.

Download

• Windows binary
• Complete source building on LibTIFF-3.7.0 (tiffprocess is in the tools directory, it is not an official and maintained contribution in the contrib folder)

Compiling on Unix/Linux:

1. execute "./configure"
2. execute "make" or "make tiffprocess",
3. and make sure that library paths are correctly set

Usage: tiffprocess input-file output-file
(without arguments, help is displayed)
Tiffprocess prompts the following inputs from the user:

• split pages or not
• source height, e.g. height of the scanned document (this is mainly for increasing robustness and can be left blank in many cases
• target height, i.e. normally height of the text area on the pages (admit a little excess since pages might not be straight)
• target width (including excess)
• resolution in DPI
• 180 degree rotation yes or no (for correct page order)
• noise reduction threshold in #pixels, this means that black pixels with less black neighbors than the specified number will be removed