Program to determine if a file is a valid image (jpg, png, gif, etc.)

KDHofAvalon · 06-09-2009, 12:57 PM

As the title says, I'm looking for a program that will go through all images in a directory tree and determine if they're valid, if they've been corrupted, and (wishful thinking) if they have the wrong extension. Does anything like that exist?

david1941 · 06-09-2009, 01:07 PM

The "file" program inspects the file contents and determines the format using a variety of tests. I've found it quite accurate in its determinations. See "man file"

harsshal · 06-09-2009, 01:17 PM

You can Google for "file type detection". Plenty of them out there. And its not that hard to create one by yourself. Look at http://en.wikipedia.org/wiki/Magic_number_(programming)

i92guboj · 06-09-2009, 01:46 PM

The identify tool from imagemagick can tell you if the file is an image or not:

Code:

$ identify jesusBrbLol.jpg && echo "Tis an image file" || echo "This is not"
jesusBrbLol.jpg JPEG 676x828 676x828+0+0 8-bit DirectClass 216kb 
Tis an image file
$ identify keysymdef.h && echo "Tis an image file" || echo "This is not"
identify: unrecognized image format `keysymdef.h' @ magick.c/ReadMAGICKImage/5571.
This is not

Checking the integrity of a file is another matter, all these tools check, as far as I know, is the header of the file.

pixellany · 06-09-2009, 01:49 PM

Quote:

Originally Posted by i92guboj

Checking the integrity of a file is another matter, all these tools check, as far as I know, is the header of the file.

And there's the real issue.....

It would be trivial to modify the header to say that the file is a particular image type. That is no guarantee that the actual data is valid.

Tell us more about the context of the question....

KDHofAvalon · 06-09-2009, 03:20 PM

Thanks for the replies.

As for why, I just recovered photos for a friend who accidentally formatted his hard drive and he want's me to delete photos that couldn't be fully recovered. There are over 10,000 so I'd really rather not go through them by hand.

farslayer · 06-09-2009, 03:48 PM

Filter 10,000 images for him.. he must be paying you very well, or a really really good friend.

not sure if you used photorec to recover the images, but their "What next After Recovery" page might be of assistance.
http://www.cgsecurity.org/wiki/After_Using_PhotoRec

KDHofAvalon · 06-09-2009, 04:54 PM

Thanks farslayer, that link had what I needed. Well sort of. It gave me an idea that works anyway.