If this is all they have then they have nothing.
The document was scanned and it was broken into layers in preparation for optical character recognition. The sections that were recognized as characters were converted to monochrome (1-bit per pixel, not grayscale). That process was not perfect which is why some of the characters were left on the color graphics layer. None of that tells whether the scanned document was taken from a real microfilm record in Hawaii and printed onto their security paper, created by Obama's friends and printed on Hawaii's security paper, or just created entirely by Obama.
Here is an image of each layer I produced when this document came out: http://www.freerepublic.com/focus/news/2711155/posts?page=997#997 .
Right now I think that we will need someone trusted to look at the microfilm of the original birth certificate to see if it matches the surrounding birth certificates (same format, same hospital name, same typeface, etc). They've had a lot of time to create paper versions, but the "evidence" this author presents proves nothing one way or the other whether this was created or a real copy of the original.
There is no OCR or optimization process that accounts for the layering, grouping, color, pixelization, and layer bit content in the whlfcolb.pdf.
I challenge anyone to reproduce it with a scan of a similar document, like a cancelled check, if they can. (and post the setting/methodology with it).