How to extract picture dimension from jpg file?

[mol]

Hi!
Has sb. any idea how to simply extract dimension of picture in pixels from jpg file?

Best regards, Marek

[danielmaximiliano]

Hello Mol:
you need to use contrib Harbour HBFimage

Greenshot_2012-01-30_17-59-13.png (60.54 KiB) Viewed 3996 times

[danielmaximiliano]

Hi:
PDF Freeimage project : http://downloads.sourceforge.net/freeim ... ge3151.pdf

PDF Complete Special Edition_2012-01-30_19-47-49.png (57.08 KiB) Viewed 3984 times

[mol]

Many thanks, Daniel, for your quick response!
Best regards, Marek

[gfilatov]

Hi!
Has sb. any idea how to simply extract dimension of picture in pixels from jpg file?

Hi Marek,

If you don't want to add the big FreeImage.dll to your project then you may take a look for the following small C-code:

Code: Select all

#pragma BEGINDUMP 

#include "windows.h"

#include "hbapi.h" 
#include "hbapifs.h"

/*
   Author: Andi Jahja <harbour@cbn.net.id>
*/
BOOL GetImageSize( const char *fn, int *x, int *y ) 
{ 
   unsigned char buf[24]; 
   long len; 

   FILE *f = hb_fopen( fn, "rb" ); 

   if (!f) 
      return FALSE; 

   fseek( f, 0, SEEK_END ); 

   len = ftell( f ); 

   fseek( f, 0, SEEK_SET ); 

   if ( len<24 ) 
   { 
      fclose( f ); 
      return FALSE; 
   } 

   // Strategy: 
   // reading GIF dimensions requires the first 10 bytes of the file 
   // reading PNG dimensions requires the first 24 bytes of the file 
   // reading JPEG dimensions requires scanning through jpeg chunks 
   // In all formats, the file is at least 24 bytes big, so we'll read 
   //that always 
   fread( buf, 1, 24, f ); 

   // For JPEGs, we need to read the first 12 bytes of each chunk. 
   // We'll read those 12 bytes at buf+2...buf+14, i.e. overwriting 
   // the existing buf. 
   if (buf[0]==0xFF && buf[1]==0xD8 && buf[2]==0xFF && buf[3]==0xE0 && 
       buf[6]=='J' && buf[7]=='F' && buf[8]=='I' && buf[9]=='F') 
   { 
     long pos = 2; 
     while ( buf[2]==0xFF ) 
     { 
       if (buf[3]==0xC0 || buf[3]==0xC1 || buf[3]==0xC2 || 
          buf[3]==0xC3 || buf[3]==0xC9 || buf[3]==0xCA || 
          buf[3]==0xCB) 
          break; 
       pos += 2+(buf[4]<<8)+buf[5]; 
       if ( pos+12>len ) 
          break; 
       fseek( f, pos, SEEK_SET ); 
       fread( buf+2, 1, 12, f); 
     } 
   } 

   fclose(f); 

   // JPEG: (first two bytes of buf are first two bytes of the jpeg 
   // file; rest of buf is the DCT frame 
   if (buf[0]==0xFF && buf[1]==0xD8 && buf[2]==0xFF) 
   { 
      *y = (buf[7]<<8) + buf[8]; 
      *x = (buf[9]<<8) + buf[10]; 
      return TRUE; 
   } 

   // GIF: first three bytes say "GIF", next three give version 
   // number. Then dimensions 
   if (buf[0]=='G' && buf[1]=='I' && buf[2]=='F') 
   { 
      *x = buf[6] + (buf[7]<<8); 
      *y = buf[8] + (buf[9]<<8); 
      return TRUE; 
   } 

   // PNG: the first frame is by definition an IHDR frame, which gives 
   // dimensions 
   if ( buf[0]==0x89 && buf[1]=='P' && buf[2]=='N' && buf[3]=='G' && 
       buf[4]==0x0D && buf[5]==0x0A && buf[6]==0x1A && buf[7]==0x0A 
       && buf[12]=='I' && buf[13]=='H' && buf[14]=='D' && 
       buf[15]=='R') 
   { 
      *x = (buf[16]<<24) + (buf[17]<<16) + (buf[18]<<8)+(buf[19]<<0); 
      *y = (buf[20]<<24) + (buf[21]<<16) + (buf[22]<<8)+(buf[23]<<0); 
      return TRUE; 
   } 

   return FALSE; 
} 

HB_FUNC( HB_GETIMAGESIZE ) 
/* 
  Syntax: HB_GETIMAGESIZE( cPicFile ) 
  Parameter: cPicFile = graphic file (JPG, GIF, PNG) 
  Return: 2 dim array -> array[1] = width, array[2] == height
*/ 
{ 
   int x = 0, y = 0; 

   GetImageSize( hb_parcx( 1 ), &x, &y ); 

   hb_reta( 2 ); 
   hb_storvni( x, -1, 1 ); 
   hb_storvni( y, -1, 2 ); 
} 

#pragma ENDDUMP

Hope that helps

[mol]

THX Grigori!
I've written my own function based on sample from internet, too!
I'll try to reduce it and I'll present it, then.

best regards, Marek

[esgici]

... I'll present it, then.

I'll wait it

Regards

--

Esgici

[mol]

Hi!
Has sb. any idea how to simply extract dimension of picture in pixels from jpg file?

Hi Marek,

If you don't want to add the big FreeImage.dll to your project then you may take a look for the following small C-code:
...
Hope that helps

I've decided to use your sample Grigori, because of its universality (3 formats) and very small code
Many thanks!


PS. While searching JPG format description, I've read that some problems with dimensions of JPG. Sometimes, this way you can read the size of miniature of JPG, not the size of image...
I'll test it

[mol]

... I'll present it, then.

I'll wait it

Regards

--

Esgici

Hi Esgici, I'll try to finish my code and add some comments, but, I think, Grigori's code looks so much interesting, so I'll use it.
regards, Marek