2009.03.14 02:04 "[Tiff] Confirming tiff integrity", by Robert Gallop

2009.03.16 07:39 "Re: [Tiff] Confirming tiff integrity", by Eric Doenges

Robert,

Is there anyway to determine (via script, executable, something I need to compile, etc) if a multipage tiff contains a solid black square without viewing the actual tiff file (what I currently have to do, hundreds of times a month). I'd think that somewhere in the tiff, there might be some clue with a normal grey scale image, vs a pure solid black square on a page. What I'd really like to be able to do is write a script, or have an application that runs against a tiff and then gives a yes or no to any solid black images contained in that multipage tiff.

You could install ImageMagik (see http://www.imagemagick.org) and use the 'identify' tool with the verbose option, which will give you a histogram of the color values in the image. Here is a sample:

Image: bitrot.tif
  Format: TIFF (Tagged Image File Format)
  Class: DirectClass
  Geometry: 576x450+0+0
  Type: Bilevel
  Endianess: MSB
  Colorspace: Gray
  Depth: 1-bit
  Channel depth:
    Gray: 1-bit
  Channel statistics:
    Gray:
      Min: 0 (0)
      Max: 1 (1)
      Mean: 0.50088 (0.50088)
      Standard deviation: 0.499999 (0.499999)
  Histogram:
    129828: (255,255,255) #FFFFFF white
    129372: ( 0, 0, 0) #000000 black
  Rendering intent: Undefined
  Resolution: 1200x1200
  Units: PixelsPerInch
  Filesize: 31.5625kb
  Interlace: None
  Background color: white
  Border color: rgb(223,223,223)
  Matte color: grey74
  Transparent color: black
  Page geometry: 576x450+0+0
  Dispose: Undefined
  Iterations: 0
  Compression: ColorBurn
  Orientation: TopLeft
  Signature:
fa9af6e337d8b28622ad507819e50474ce792b086b0e98d0b9182c72cbdf8045
  Tiff:rows-per-strip: 450
  Tainted: False
  Version: ImageMagick 6.3.5 05/27/08 Q16 http://www.imagemagick.org

If the image is completely black, you should have only one entry in the histogram (if invalid images can contain some white pixels, you could calculate what percentage of the image is black by comparing the number of black pixels with the total number of pixels).

With kind regards,
Eric
--
Dr. Eric Dönges doenges@mvtec.com

MVTec Software GmbH | Neherstr. 1 | 81675 München  | Germany

www.mvtec.com | Tel: +49 89 457695-0 | Fax: +49 89 457695-55

Geschäftsführer: Dr. Wolfgang Eckstein, Dr. Olaf Munkelt
Amtsgericht München HRB 114695