Tesseract OCR for Kannada

Everywhere I go I field requests like this one, about a good Kannada OCR system for a research project. However after looking around for a while I found that there isn't a good solution. Once in a while people report it works well with this language (or that) but there is always some nagging problems with complex Indian language scripts. Can anyone tell how well the Tesseract OCR system perform with Kannada language.

Read More

Handling CHM files in GNU/Debian Linux

In a perfect world all the OCR files would come in djVu format but world isn’t perfect, so this morning I had read this 300 page CHM compiled HTML Help (insert joke here)file and for no apparent reason my CHM viewer won’t display the file beyond page 50. Bummer!.

Don't get me wrong but there some great programs to handle CHM files on gnu/Linux OS. And I’ll mention some of them here in this post. A more detailed specifications of CHM format. is available if you are interested.

Graphical CHM Viewers

Well, The brief list doesn’t attempt to be complete but it gives you an idea of range of tools available to a gnu/Linux user.

• Kchm


• xCHM


• GnoCHM



• CHMsee


• Archmage


• DisplayCHM (contributed by Rajasekaran)


• CHM Firefox addon (contributed by Rajasekaran)



• (if I missed any, do leave a comment with a note)

View a CHM file on a Web Browser

If you like to read using your favorite web browsers instead of a Graphical X based CHM viewer, then there is chm_http in libchm-dev package from the CHMLib project . Also Archmage package has http-server and or if you prefer Apache Web Server instead of standalone webservers then there is an extension mod_chm too .

Converting and De-compiling CHM files

Of late I saw chm2pdf project offering to covert CHM files directly into PDF. I should try it sometime, but for now I will just de-compile the CHM file and get done with it.

So, if you don’t have libchm-dev package installed, then do this.

sudo apt-get install libchm-dev

This package has “extract_chmLib” program which will extract and place the files in a given directory.

extract_chmLib Foo.chm Foo-dir/

That it, now you can use your web browser to read the contents at your will. It is a quick and dirty way, but its much better than using the dreaded CHM files anyway.

Compiling CHM files ?

Sometime after this entry was posted, I got people asking me how to create CHM files under gnu/Linux .

According to Bruce Byfield in two year old article on FOSS help authoring tools


Numerous projects for cross-platform or GNU/Linux-based tools for producing .chm files have been announced, but most of them are in the planning stage, and have released little or no code. The only two that seem polished enough to use are AurigaDoc and DocBook two XML-based solutions that are well-known in other contexts.

Read More