[ASSP_OCR] imagemagick convert problem?

[bku09]

hi,

ASSP_OCR plugin always fails converting images with error code -1 :

Code: Select all

ASSP_OCR::get_ocr(), imagemagick convert problem? "/usr/bin/convert" "/assp2/tmp/ocr_pl13280166209144~WRD000.jpg" -compress none -colorspace rgb -contrast "/assp2/tmp/.tesseract_x_temp_1328016620437.tif", -1


but if I extract the above jpg image and convert it manually via cmdline, it works without problems (exit code 0) :

Code: Select all

# /usr/bin/convert "/assp2/tmp/ocr_pl13280166209144~WRD000.jpg" -compress none -colorspace rgb -contrast "/assp2/tmp/.tesseract_x_temp_1328016620437.tif"
# echo $?
# 0


tested with:
Ubuntu 10.04.3 LTS
Assp 2.1.1(11364)
ASSP_OCR,v 2.02 2012/01/06 (older versions doesen't work either)
ImageMagick 6.5.7-8 (Ubuntu shipped)
ImageMagick 6.7.4-1 (self compiled)

any ideas?

thanks,
Bernd

[schefczyk]

Dear All,

I have a similar problem and I would very much welcome feedback to the post above.

My observation is that attached pdf-images are processed by assp_ocr with no problem at all. Results in terms of conversion and extraction are retained in the relevant tmp-directory for a while in the form of .tif and .tif.txt files. However, jpg and png files in the body of an e-mail, such as logos with little text, generate the convert problem indicated in the previous post. I suspect that DoPDF ... does work while DoImageASSP_OCR does not work for some reason.

I did use assp V2 for a while on Centos to switch to Ubuntu/Xubuntu this earlier this month given Perl requirements. I use ASSP version 2.2.1(13020) on Xubuntu 12.04 on a virtual machine with (currently just) two cpu cores.

Regards,

Michael