In general I've been quite successful in extracting images by navigating the iText (http://itextpdf.com/) data-structure, identifying the encoding and applying the appropriate decoder to the PdfStream.

I have a slightly more complex situation where it appears the images are in a PDF form encoded with FLATEDECODE

So I decode the FLATEDECODE data and end up with the following:

/GS5 gs
q
0 0 1 1 re
W n
/GS4 gs
0 J 0 j 4 M []0 d 0.0284 w
0 0 0 0.7 K
0.9539 0.5332 m
0.9891 0.5332 l
0.9891 0.5896 l
0.9697 0.6397 l
0.4126 0.6397 l
0.4126 0.337 l
0.9697 0.337 l
0.9912 0.3718 l
0.9912 0.4193 l
0.6447 0.4193 l
0.6792 0.5332 l
0.7494 0.5332 l
0.7494 0.4957 l
0.7814 0.4957 l
0.7814 0.545 l
0.8219 0.545 l
0.8219 0.4994 l
0.8557 0.4994 l
0.8557 0.5488 l
0.8982 0.5488 l
0.8982 0.4945 l
0.9315 0.4945 l
0.931 0.5327 l
s
0.7 0 0 0.45 k
0.0118 0.1781 m
0.0757 0.0868 0.1632 0.0305 0.2595 0.0305 c
0.356 0.0305 0.4433 0.0868 0.5073 0.1779 c
0.5073 0.8219 l
0.4433 0.9132 0.3559 0.9695 0.2595 0.9695 c
0.1631 0.9695 0.0757 0.9132 0.0118 0.822 c
0.0118 0.1842 l
f
0.0379 w
0 0 0 1 K
0.0118 0.1781 m
0.0757 0.0868 0.1632 0.0305 0.2595 0.0305 c
0.356 0.0305 0.4433 0.0868 0.5073 0.1779 c
0.5073 0.8219 l
0.4433 0.9132 0.3559 0.9695 0.2595 0.9695 c
0.1631 0.9695 0.0757 0.9132 0.0118 0.822 c
0.0118 0.1842 l
s
0 0 0 0.7 K
0.9776 0.5242 m
0.9776 0.4343 l
S
0 0 0 1 k
0.3677 0.5177 m
0.3677 0.3633 0.3193 0.2382 0.2597 0.2382 c
0.2001 0.2382 0.1517 0.3633 0.1517 0.5177 c
0.1517 0.672 0.2001 0.7972 0.2597 0.7972 c
0.3193 0.7972 0.3677 0.672 0.3677 0.5177 c
f
0.0265 w
0 0 0 0 K
0.3677 0.5177 m
0.3677 0.3633 0.3193 0.2382 0.2597 0.2382 c
0.2001 0.2382 0.1517 0.3633 0.1517 0.5177 c
0.1517 0.672 0.2001 0.7972 0.2597 0.7972 c
0.3193 0.7972 0.3677 0.672 0.3677 0.5177 c
s
0 0 0 0 k
0.2107 0.4173 m
0.2107 0.5479 l
0.3071 0.5479 l
0.3071 0.4173 l
0.3071 0.4152 l
0.3071 0.355 0.2855 0.3061 0.2589 0.3061 c
0.2323 0.3061 0.2107 0.355 0.2107 0.4152 c
f
0.231 0.5074 m
0.231 0.6125 l
0.231 0.6569 0.2449 0.693 0.2621 0.693 c
0.2793 0.693 0.2932 0.6569 0.2932 0.6125 c
0.3054 0.6125 l
0.3054 0.6743 0.286 0.7244 0.2621 0.7244 c
0.2382 0.7244 0.2188 0.6743 0.2188 0.6125 c
0.2188 0.5074 l
f
0 0 0 1 k
0.2733 0.4696 m
0.2733 0.4487 0.2667 0.4319 0.2587 0.4319 c
0.2506 0.4319 0.2441 0.4487 0.2441 0.4696 c
0.2441 0.4905 0.2506 0.5074 0.2587 0.5074 c
0.2667 0.5074 0.2733 0.4905 0.2733 0.4696 c
f
0.2695 0.3512 m
0.2478 0.3512 l
0.2516 0.4706 l
0.2664 0.4706 l
f
Q


1. Exactly what format is the above, is it raw ghost-script because I can't seem to read it with anything (gsview,acrobat etc)?

2. If it is an EPS (maybe it's something else?) what do I need to do to turn this into and viewable image?

Can you please be a little bit more verbose I do not understand?

If I am wrong, and you have read the document, perhaps we can find you find out where you have gone wrong, because this is not an image. What makes you believe it is an image?

...is this the one?

http://www.adobe.com/devnet/pdf/pdf_reference.html

It's also likely that I'm going completely down the wrong path to so I'll explain my problem in more detail.

I am try ing to extract FPO low res images that appear in a PDF document (those appearing with OPI comments)

For this particular document (I chose this one a random from the internet) I have problems:

http://media.wiley.com/product_data/excerpt/18/EHEP0000/EHEP000018-1.pdf

Now I know the image exists because I can see it on the page so the image is there somewhere (the low res one). Note the F PdfName refers to the high res image (which I don't care about for now)

This is the object which is identified (for page 3) as the OPI reference for the image. Normally it would have a reference to an XObject (which is the low res image) but I can't find it so I assumed it was in a form connected to this obj?

^M25 0 obj<>>>/Subtype/Form/Length 70/Filter/FlateDecode/Name/Fm20/Matrix[1 0 0 1 0 0]/Resources<>/ProcSet[/PDF]/ExtGState<>>>/Type/XObject/BBox[0 0 1332.0 783.0]/FormType 1>>stream