When you say "parse an image", what exactly do you want to do?
The PartsList sample will show you which part your image is in. http://dev.plutext.org/trac/docx4j/brow ... sList.java
Once you have the part, you can for example save the image to a file. But that might not be what you want to do.
Each image part will be referenced from somewhere (eg document.xml or a header or footer), using a relationship id.
You can change the relationship id to point to a different image.
When docx4j is used for HTML or PDF export, it uses the following package to handle images (nb, this was cleaned up substantially following the release of 2.3.0). http://dev.plutext.org/trac/docx4j/brow ... del/images
cheers .. Jason