How to convert vector in PDFs to image?

[wmr]

Hi,

I am using mobipocket to convert a textbook PDF file to .mobi. Some graphics/formulas when wrong after conversion.

After searching the web for a while, I found K2pdfopt but not very satisfied. It turns everything into images.

I need a software to convert every graphic and formula in PDF to image, in which I can manually select graphics and formulas in that file to convert. Thus I can convert it to .mobi without trouble.


Cheers.

[Toxaris]

What I usually do is open the PDF page I want in GIMP, but there are more possibilities like Imagemagick. However, that would result into bitmaps. For formula's SVG or MathML would be best, but that will require recreating the formulas. Downside for SVG is that it is not supported in mobi AFAK and MathML is only supposed to be supported in ePUB3.

If I am wrong and SVG is supported in mobi, you can try the codecogs site to convert a formula to SVG. If not, you would be best of with a PNG. If you optimize the PNG, they can be rather small in size.

[Tex2002ans]

How many vector graphics/formulas are we talking here? It can get quite labor intensive if you have to do a lot of them without access to the original source documents. Not to mention, you probably would have to manually rename the filenames to organize them neatly as well.

I would take a look at "Step 3: Inkscape" of my "Tutorial: Formulas to PNG" topic:

https://www.mobileread.com/forums/sho...d.php?t=223254

I have a bunch of examples of formulas I generated using that method in the tutorial. And here is a vector graph that was originally designed from InDesign, I pulled out of the PDF using Inkscape, and exported to PNG:

https://www.mobileread.com/forums/sho...5&postcount=26

You can easily follow the same steps in Inkscape, and export the formulas/graphs from PDF -> SVG instead of PNG if that is what you prefer.

Originally Posted by Toxaris

Downside for SVG is that it is not supported in mobi AFAK and MathML is only supposed to be supported in ePUB3.

SVG is only supported in KF8, so if you wanted your MOBI to work on older Kindles as well, you would need to create image fallbacks anyway.

If this is a personal project, and you only have a newer Kindle device, then you should be ok with SVG.

Originally Posted by Toxaris

If you optimize the PNG, they can be rather small in size.

Yep, if you Index the PNGs, and reduce the colors, you can get the filesize of formulas QUITE small.

For example, the latest journal I did, 81 formulas at 350 DPI = 328KBs after optimization. (Samples are attached in the topic).

I just have two words of advice: AVOID JPG.

[markom]

Originally Posted by wmr

Hi,

...

I need a software to convert every graphic and formula in PDF to image, in which I can manually select graphics and formulas in that file to convert. Thus I can convert it to .mobi without trouble.


Cheers.

ABBYY FineReader.

You can select any part of text with formulas and then by right-clicking on it change its "area type" to picture and then (after you passed through whole pdf) save pdf as "text and pictures only" or as epub(with "keep pictures" box checked and dpi for pictures as desired e.g. 300 dpi or best quality).

Pictures and tables shall be recognized automatically (visible as red and blue area) but you should probably change tables (blue) to pictures (red) too.

I don't know if it is possible to automatize it though, like it is possible to make Abbyy automatically search for formulas and its special signs and teach it desired OCR behaviour, to automatically correct spotted errors.