Extract Images from PDF

The size filter is the point

Here is what most extractors do not tell you: a PDF is usually full of images that are not pictures. Transparent spacers used for layout. Bullet glyphs. Slivers of a logo, sliced up by whatever produced the file. One-pixel gradient strips stretched across a header. A design-tool export can easily contain sixty of them.

Extract without a filter and you get all of it — a ZIP of dozens of files, most of them useless, and the three photographs you actually wanted buried among them.

The minimum image size setting cuts that off at the source. 100px is the default and removes essentially all layout debris while keeping anything that is genuinely a picture. 200px narrows it to photographs and large diagrams — the right choice for a brochure or catalogue where you only want the product shots. 50px opens it up for when you are specifically hunting something small, like a logo or an icon you know is in there.

Two different things people mean by this

Worth separating, because it is the main source of disappointment with tools in this category.

Extracting embedded images — what this does — reaches into the PDF and pulls out the image files that were placed into it. You get the photographs as photographs, at whatever size and quality they were embedded.

Turning pages into images is a different job: each page becomes one picture, text and layout and all. That is what you want for a preview, a thumbnail, or posting a page somewhere that will not display a PDF — and it is a separate tool.

If you are not sure which you need, ask what you want to end up holding: the photo from page four, or a picture of page four.

What cannot be extracted

Vector graphics. A chart, an icon set or a diagram drawn as vectors is not an image inside the PDF at all — it is a set of drawing instructions, lines and curves and fills executed when the page renders. There is no file in there to pull out, so it will not appear in the results however you set the filter. To get a copy of one, render the page as an image instead and crop it, or go back to whatever produced the chart.

And if direct extraction cannot read the images in a particular file, the tool falls back to rendering each page as an image. That is a different result and worth recognising: if your ZIP contains full pages rather than individual pictures, that is the fallback having run, not a filter problem.

What quality to expect

You get the image exactly as it was embedded, which may be less than the original. A PDF made for print usually carries images at full resolution, so extraction gives you something genuinely usable. A PDF made for the web or email has usually had its images downsampled and re-compressed to keep the file small, so what comes out is the reduced version — extraction recovers what is in the file, not what the file was made from.

So if the extracted images look softer than you hoped, that is the document rather than the tool, and the original source is the only place a better copy exists.

A note on using them

Being able to extract an image is not the same as being free to reuse it. Photographs in brochures, reports and presentations are usually licensed or owned by someone, and a stock photo inside a PDF is still a stock photo. Extracting from your own documents, or to recover your own assets, is what this is for.

Nothing is uploaded, incidentally, which matters more here than for most conversions — the PDFs people pull images out of include scanned identity documents, medical reports and unreleased design work.

Related tools

PDF to JPG is the other meaning: it turns each page into an image, which is what you want for previews and thumbnails. PDF to TXT Converter gets the words out instead, and Compress PDF shrinks a PDF whose images are making it huge.