MobileRead Forums - View Single Post - Remove background images from pdfs? perhaps all images?

MarjaE · 10-17-2018, 12:37 AM

Hi,

I have a few pdfs I can't read because background images obscure the text. I don't expect any solution for scanned pdfs, but I've tried to find one for pdf-born-pdfs, and been beset with bugs.

In Ghostscript, I've tried:

gs -sDEVICE=pdfwrite -dFILTERIMAGE -dFILTERVECTOR -dCompatibilityLevel=1.4 -dNOPAUSE -dQUIET -dBATCH -sOutputFile=[output.pdf] [input.pdf]

I don't just lose raster images and vector images, I lose about half the text too. And quick checks confirm it wasn't raster images of text.

I've also tried a 2-step process with:

gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 -dNOPAUSE -dQUIET -dBATCH -sOutputFile=[output.pdf] [input.pdf]

*and then*

gs -sDEVICE=pdfwrite -dFILTERIMAGE -dFILTERVECTOR -dCompatibilityLevel=1.4 -dNOPAUSE -dQUIET -dBATCH -sOutputFile=[output.pdf] [input.pdf]

Now I lose about one-twentieth of the text instead of half, but that's still too much. I usually end up with the lower left corner of the page blown up to fill the whole page.

I've tried using mutool clean -d -l -g, or cpdf with specified page sizes (and -blacktext to avoid white text on white backgrounds), or ghostscript with specified page sizes, but none of these solve the problem.

Any suggestions?

10-17-2018, 12:37 AM	#1
MarjaE Guru Posts: 939 Karma: 53902736 Join Date: Jun 2015 Device: multiple	Remove background images from pdfs? perhaps all images? Hi, I have a few pdfs I can't read because background images obscure the text. I don't expect any solution for scanned pdfs, but I've tried to find one for pdf-born-pdfs, and been beset with bugs. In Ghostscript, I've tried: gs -sDEVICE=pdfwrite -dFILTERIMAGE -dFILTERVECTOR -dCompatibilityLevel=1.4 -dNOPAUSE -dQUIET -dBATCH -sOutputFile=[output.pdf] [input.pdf] I don't just lose raster images and vector images, I lose about half the text too. And quick checks confirm it wasn't raster images of text. I've also tried a 2-step process with: gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 -dNOPAUSE -dQUIET -dBATCH -sOutputFile=[output.pdf] [input.pdf] and then gs -sDEVICE=pdfwrite -dFILTERIMAGE -dFILTERVECTOR -dCompatibilityLevel=1.4 -dNOPAUSE -dQUIET -dBATCH -sOutputFile=[output.pdf] [input.pdf] Now I lose about one-twentieth of the text instead of half, but that's still too much. I usually end up with the lower left corner of the page blown up to fill the whole page. I've tried using mutool clean -d -l -g, or cpdf with specified page sizes (and -blacktext to avoid white text on white backgrounds), or ghostscript with specified page sizes, but none of these solve the problem. Any suggestions?