Best way to copy and modify pictures from sources like (pdf, web) so that the modified image has high quality?

Question

I am using TeXMaker (Latex) to write my PhD thesis, I have a lot of thesis examples (extension .PDF, .tex) I want a help concern the copy and modification of pictures from sources (web, book, thesis, reports, PDF), for example when I locked to theses thesis In most of them, they used pictures or photos from other resources and they modify them to get their needs preserving the high quality of the modified image (in the produced pdf), the way that I am using to copy, modify, and save my images is as follows:

Using the PDF selection tool to make a snapshot from the source (if the image is in the web I use Snagit tool to make a screenshot

and save the image in one of the various format allowed by this tool).

Copy the selected graphics (image) to a word file.

Make my necessary changes.

Using Snagit tool I make a screenshot and save the image (.jpg, pdf, png, ....)

This the way that I am using to create images for my PHD thesis, but when I include these images in Latex the images in the

produced PDF look bad (I am using Foxit reader as a pdf viewer).

Is there any way or tool should I use to create good modified images for my thesis?

For example the following image, I created it using the procedure mentioned before, the image it looks nice when I open it using a photo viewer (windows photo viewer), but for some reason, it looks bad when I create my PDF file.

\documentclass{book}
\usepackage{graphics}
\begin{document}

\begin{figure}[htbp]
\centering
\includegraphics{PHDImages/Neuron.jpg}
\caption{Neuron}
\end{figure} 
\end{document} 

Answer 1

Adobe Acrobat Reader has a feature that allows you to copy content from a PDF that, if I remember correctly, preserves the image quality. This would be a good option for raster images.

For vector illustrations I have successfully used Inkscape in the past. Import the pdf file in Inkscape and select the page you want to open as an SVG file. Then use Inkscape's tools to extract the illustration which can then be saved as a vector EPS file.

Answer 2

You obviously use screenshots (or images with the same resolution, you copy them from web pages) because the used data is available most easily this way (no comment here about how legal it's to grab something from web an republish it, that will be only your problem).

Getting original sharp images instead of screenshots are obviously also out of question for some reason. Otherwise you had used them and no problem would exist.

Bitmap images suffer every time they are resized or saved as JPGs. Screenshots generally have no quality reserve, they become inferior when resized and that's your poblem.

Your options:

• do not make any rescaling, then you get the original screenshot sharpness (assuming you have no downgrading selected in the PDF making settings), but definitely not something printable. Also do not save as JPG, make PNGs, if possible, JPGs have compression losses.
• redraw the image in high resolution or vector format including the drawing and the texts
• trace the image in Inkscape - If you are lucky and carefully removed texts and other non-autotraceable things, you can get an usable vector image with reasonably minimal redrawing. Your included image is a hopeless case, it has far too low resolution (468 x 514 pixels)
• use Smilla or other photo enlargening software to increase the bitmap resolution to minimum print quality value, say 300dpi. (=enlarge your example to about 400% or more) Those programs are guite good to guess the needed bits in simple cases.

Enlargening do not make automatic tracing to vector better, it only makes the annoying wrong tracing result details smaller but as harmful looking.

Here is your image in 400% pixel dimensions after proper enlargening:

If you wipe off all texts before enlargening, enlarge and finally rewrite the texts you probably have enough resolution. No enlargening software unfortunately can guess the texts right from a screenshot. That's because displaying text as low res screen image needs heavy processing to fool the eye to see it readable.

JPGs often have plenty of compression artefacts. Your image is not an exception. The artefacts are the complex light pattern around all edges. They should be cleared in a bitmap image editor such as GIMP or Paint.NET (both freeware) before using the image to anything.

Here is a big size screenshot of your original image. The light grey stray pixels inside my red curve are all a small part of the JPG atrefact pattern. Just in this case they are easily turned to white with the curves tool in a bitmap image editor.

Of course you lose also image details, but in this lucky case they are mostly texts, which you should anyway rewrite. Here's how curves affect (Photoshop):

Artefacts make enlargening results and automatic tracings, too, substantially worse. It needs much manual work to clean the images, automatic methods found in many bitmap image editors and noise removal software often easily wipe off also image details.

Instruction includegraphics accepts PDFs. Assuming TeXMaker do not spoil included PDFs, you should compose your images in a program which can insert texts and auxiliary drawings as vector data and make them PDFs. Free example: Inkscape.

Here's a screenshot from PDF reader. I placed the 400% version of your image to Inkscape, added random texts and graphics and saved as PDF.

"Haven't that font" -problem do not exist if you select option convert texts to curves ON as you save as PDF

Unfortunately I haven't any deep knowledge of TexMaker. I cannot tell what possiblities there's to parametrize the making of the final PDF. You should find a way to prevent any bitmap image downscaling and extended JPG commpression. If you happen to use actually some PDF printer software - such as Cute PDF or Adobe PDF to make the PDF, check it's documentation.

Inkscape knows LaTeX, you probably have more workflow options than I can imagine.

Get local help - a person who knows how to get the most out of available material and software and can guide you interactively. We have nothing but one low res image and a complaint "this is fine, but the PDF is bad, I want good PDF!"