Thumbnail removal with Sanselan

[Marko Mrkus <ma...@bumblehood.com>]

Hi!

As I see Sanselan provides a very easy way how to read and modify EXIF
headers in JPEG files.

Can you, please, tell me is it capable of removing thumbnail from JPEG and
how to approach this problem?

 

PS. keep up the good work :-)

 

Thanks,

Marko

Re: Thumbnail removal with Sanselan

[Charles Matthew Chen <ch...@gmail.com>]

Hi Mark -

   Thanks for your interest in the project.

   This is an interesting idea for a new feature.  One feature I've
thought of adding for a long time would be a format-agnostic "strip"
function that would remove everything except the image data from an
image - ie. metadata, comments, thumbnails, etc.  This would certainly
do what you need, but I think you're looking for something more
focused.

   Removing just the Exif thubnails from Jpegs is actually fairly
simple.  Jpeg stores its EXIF metadata in a TIFF-structured segment.
TIFF data is arranged in a series of directories.  The thumbnail (if
present) is data is in a contained is a dedicated directory - removing
a thumbnail is thus a question of finding that directory and removing
it if present.

   It will require a fairly simple patch.  Here's how it will go:

1. Your code will be a variation on the jpeg metadata example code:

https://svn.apache.org/repos/asf/incubator/sanselan/trunk/src/test/java/org/apache/sanselan/sampleUsage/WriteExifMetadataExample.java

...
outputSet = exif.getOutputSet();
TiffOutputDirectory 	thumbnailDirectory = outputSet.getExifThumbnailDirectory();
if(null != thumbnailDirectory)
   outputSet.removeDirectory(thumbnailDirectory);
else
   // if the jpeg's exif metadata doesn't contain a thumbnail, should
we return and not bother updating the image?
// continue updating the image as in the example.
...

2. Second, we need to add a couple of methods to TiffOutputSet.java.

https://svn.apache.org/repos/asf/incubator/sanselan/trunk/src/main/java/org/apache/sanselan/formats/tiff/write/TiffOutputSet.java

...
// new  code
	public TiffOutputDirectory getExifThumbnailDirectory()
	{
		// I can't recall if the proper value here is DIRECTORY_TYPE_THUMBNAIL
		// or DIRECTORY_TYPE_SUB.
		//
		// see: https://svn.apache.org/repos/asf/incubator/sanselan/trunk/src/main/java/org/apache/sanselan/formats/tiff/constants/TiffDirectoryConstants.java
		return findDirectory(DIRECTORY_TYPE_THUMBNAIL);
	}
	public void removeDirectory(TiffOutputDirectory directory)
			throws ImageWriteException
	{
		if(!directories.contains(directory))
			throw new ImageWriteException(
					"Output set does not include that directory.");
		directories.remove(directory);
	}
// end new code

   I haven't compiled or tested this code, so it may have some obvious
mistakes in it.  If you get this working, please consider contributing
a patch to the project by:
a) opening a JIRA ticket https://issues.apache.org/jira/browse/SANSELAN
b) attaching your code to the ticket.

Thanks,
   Charles


On Tue, Feb 17, 2009 at 5:41 AM, Marko Mrkus <ma...@bumblehood.com> wrote:
> Hi!
>
> As I see Sanselan provides a very easy way how to read and modify EXIF
> headers in JPEG files.
>
> Can you, please, tell me is it capable of removing thumbnail from JPEG and
> how to approach this problem?
>
>
>
> PS. keep up the good work :-)
>
>
>
> Thanks,
>
> Marko
>
>