OSD Not working --psm=0,1,12

[abir26]

Environment

• Tesseract Version: Tesseract v4.0.0-beta.1-77
• Commit Number: g8182
• Platform: ubuntu 16.04
• Leptonica Version: Leptonica-1.75.3

Current Behavior:

I get this error "Error: Illegal Parameter specification!
"Fatal error encountered!" == NULL:Error:Assert failed:in file globaloc.cpp, line 75" or an empty output whenever I use the OSD (--psm=0,1,12). Also when I tried to extract Orientation, Direction and Skew angle I get 0 for all of them no matter what picture I use.

This is the picture I used:

In command line I get this:

I tested it with the c++ code "Orientation and script detection (OSD) example" provided in the wiki:
tesseract::TessBaseAPI *api = new tesseract::TessBaseAPI(); api->Init(dataPath, "eng"); api->SetPageSegMode(tesseract::PSM_AUTO_OSD); api->SetVariable("tessedit_ocr_engine_mode","0"); api->SetVariable("tessedit_enable_dict_correction","1"); Pix *image = pixRead(imagePath); api->SetImage(image); tesseract::Orientation orientation; tesseract::WritingDirection direction; tesseract::TextlineOrder order; float deskew_angle; api->Recognize(0); tesseract::PageIterator* it = api->AnalyseLayout(); it->Orientation(&orientation, &direction, &order, &deskew_angle); printf("Orientation: %d;\nWritingDirection: %d\nTextlineOrder: %d\n Deskew angle: %.4f\n", orientation, direction, order, deskew_angle); // Get OCR result outText = api->GetUTF8Text(); meanConf=api->MeanTextConf();
I get this result :

And I tried it with a java code with the skewed image and Euronext.tif, I got this error:

[amitdo]

--psm 0 not --psm=0
Same for oem.

Also see:
#1167 (comment)

[Shreeshrii]

$tesseract -v
tesseract 4.0.0-beta.1-64-gd284
 leptonica-1.76.0
  libjpeg 8d (libjpeg-turbo 1.3.0) : libpng 1.2.50 : libtiff 4.0.3 : zlib 1.2.8 : libopenjp2 2.3.0
 Found AVX
 Found SSE

1. Use -l osd (not eng)

2. For command line, the following works (modify paths for your setup)

$tesseract devatest-rotated-270.png stdout  --tessdata-dir /mnt/c/Users/User/shree/tessdata_fast --psm 0 --oem 0 -l osd

Page number: 0
Orientation in degrees: 270
Rotate: 90
Orientation confidence: 7.07
Script: Devanagari
Script confidence: inf

3. For api example, see the following

#include <tesseract/baseapi.h>
#include <leptonica/allheaders.h>
int main()
{
    const char* inputfile = "/mnt/c/Users/User/shree/tesseract-HEAD/testing/devatest-rotated-270.png";
    PIX *image = pixRead(inputfile);
    tesseract::TessBaseAPI *api = new tesseract::TessBaseAPI();
    api->Init(NULL, "osd");
// Get OSD - old example code
    tesseract::Orientation orientation;
    tesseract::WritingDirection direction;
    tesseract::TextlineOrder order;
    float deskew_angle;
    api->SetPageSegMode(tesseract::PSM_AUTO_OSD);
    api->SetImage(image);
    api->Recognize(0);
    tesseract::PageIterator* it =  api->AnalyseLayout();
    it->Orientation(&orientation, &direction, &order, &deskew_angle);
    printf("Orientation: %d;\nWritingDirection: %d\nTextlineOrder: %d\n" \
    "Deskew angle: %.4f\n",
    orientation, direction, order, deskew_angle);
//Get OSD - new code
    int orient_deg;
    float orient_conf;
    const char* script_name;
    float script_conf;
    api->DetectOrientationScript(&orient_deg, &orient_conf, &script_name, &script_conf);
    printf("************\n Orientation in degrees: %d\n Orientation confidence: %.2f\n"
    " Script: %s\n Script confidence: %.2f\n",
    orient_deg, orient_conf,
    script_name, script_conf);
    // Destroy used object and release memory
    api->End();
    pixDestroy(&image);
    return 0;
}

Console output

Detected 127 diacritics
Orientation: 3;
WritingDirection: 0
TextlineOrder: 2
Deskew angle: -0.0071
************
 Orientation in degrees: 270
 Orientation confidence: 7.07
 Script: Devanagari
 Script confidence: inf

[Shreeshrii]

Re: "Error: Illegal Parameter specification!

See #1010

we have no problem for the tesseract executable or the training programs which are provided by Tesseract. Nor will external software have a problem as long as it does not set LC_NUMERIC.
Maybe Java uses the environment settings to set LC_NUMERIC internally. That would explain the reported problem.

So, far java, python etc

Locale should rather be set explicitly. The workaround is to set LC_NUMERIC=en_US.UTF-8.

[abir26]

ok now it works when I used --psm 1 --oem 0 for the images :phototest-rotated-R/L/180.png . However it doesn't work with the image above. same for the C++ code.

[amitdo]

Paste the full command.

it doesn't work with the image above

What do you mean by 'it doesn't work'? What's the output?

[abir26]

I thought Tesseract does the image processing operations internally so it will determine the skew angle, straighten and do the OCR on the processed image. But here I get the orientation in degrees and the Deskew angle equal to 0 for the image Euronext_skew.tif and obviously couldn't recognize the image content.


api example code:

Anyway, I will just need to improve the quality of the OCR input externally in my project.
Thank you for your help.

[amitdo]

tesseract --tessdata-dir ~/tesseract/tessdata/fast ~/Downloads/skwedlat.png - -l Latin --psm 6

The (quick) [brown] {fox} jumps!
Over the $43,456.78 <lazy? #90 dog
& duck/goos¢. as 12.5% of g-mail
from aspammer@website- cu is spam.
Der schnelle” braune Fuchs springt
über den faulen Hund. Le renard brun
«rapide» saute par-dessus ie chien
paresseux. La volpe marrone rapida
csalta sopta i] cane pigro. E1 zorro
marrón rápido salta sobte el perro
perezoso. A raposa marrom rápida
salta sobte o cão preguiçoso.

tesseract --tessdata-dir ~/tesseract/tessdata/best ~/Downloads/skwedlat.png - -l Latin --psm 6

The (quick) [brown] £fox}3 jumps!
Over the $43,456.78 <1azy7 490 dog
& duck/g00S¢> as 12.5% of E-mail
from aspammer@website-Cat is spam.
Der schnelle” braune Fuchs springt
über den faulen Hund. Le renard bruB
«rapide» saute par-dessus le chien
paresseu*X-. La volpe marrone rapida
salta sopra jì cane pigro- E1 zorro
marrón rápido salta sobre el perro
perezos0. A raposa marrom rápida
salta sobre o cão preguiç0s0-