gocr

I want to read with gocr images of text that i grab from websites using grab.

grab can output them as jpeg, png or ILBM

I am not sure what files is should download ? where to find them.
gocr.tcl is not recogised as an executable, so i just tried using gocr from shell. For a png file a pngtopnm file seems to be needed? For a jpeg file djpeg. Other files are mentioned too gzip, etc. Where can i find them?

@JosDuchIt

Hmm, I ported gocr but I can't remember how to use it. I've just scanned through the docs and it looks like you need your source file to be PNM. You can use jpegtopnm to achieve this.
jpegtopnm file_to_convert.jpg >output.pnm (realistically it will output a PPM)
You can get jpegtopnm by downloading the NetPBM package.
You may be able to feed a jpeg straight into gocr when that is installed, there's a mention of that but I have no idea whether it works.

Then run it through gocr using something like
gocr output.pnm >output.txt

The character recognition wasn't good enough for my needs, hopefully you'll have better luck with it.

@Chris
Thanks for the hints.
i was able te make it work from shell using jpegtopnm or pngtopnm first then using gocr. Results were not brilliant, but on the PC ABBYY fineprint 6.0 was not better, although this SW allowed me to cut out a block to be read.

I am looking now for a simple SW permitting to just cut out a rectangular part of the image "CutOut This Image X, Y, Width, Height" from shell or using arexx? Any suggestion?

@JosDuchIt
Try PropperCropper from OS4Depot. It saves as JPEG so you may need another conversion to use the cropped image.

@xenic

Quote:
With ImageMagick you should be able to crop it and save it in any format you like. The port is quite old and I'm not sure if it will run under the latest OS 4.1, though:

http://os4depot.net/index.php?functio ... /edit/imagemagick-bin.lha

Here's a tutorial on the crop function:

http://www.imagemagick.org/Usage/crop/

Varthall

I used to have a good OCR that came with a Migraph handy scanner. It worked quite well using a "learn" feature. But I think that the bundled version only worked together with the handy scanner...
If I find it on an old CD backup, I can try to see if one can make it work...

Now, my version was limited (only works together with the scanner), but here you will find the complete version :


(dunnow if it is legal)


[Edit by afxgroup - Removed link to illegal software]
yes.. it is..

yess, it works
You have to transform your images to monochrome IFF.
I don't know why, but the output from BME (the graphics editor of Pagestream) is not recognized, but if you re-save it through PPaint, it works!!