What counts as an image in a PDF
A PDF page is built from text, vector graphics and images. Only images are made of pixels, and those are what this tool finds: photos, scanned pages, screenshots and any other raster picture the page paints. It reads the drawing instructions of every page with pdf.js, the open source PDF engine from Mozilla, and lists each picture with the pages it appears on, its pixel size and how the PDF stores it.
- Images: pictures stored as separate objects in the file. Most photos and scans are stored this way.
- Inline images: small pictures written straight into the page. The PDF standard reserves them for images of 4 KB or less.
- Stencil masks: one-bit shapes painted in a single color, such as a scanned signature or stamp. They are saved as black on a transparent background.
Vector graphics are not images. Logos, charts, diagrams and icons are often drawn with lines and curves, which have no pixels to extract. To capture those, save the page with PDF to JPG and crop what you need. Some programs also cut a big picture into strips or tiles, which then show up as several images, and many images carry a soft mask for transparency, which the PNG file keeps.
Original JPEG or lossless PNG
A photo inside a PDF is usually stored as JPEG data, marked with the DCTDecode filter. When that data can stand on its own, the tool saves it as a .jpg file without decoding or recompressing it, so you get the exact bytes that were put into the PDF. Everything else is decoded by pdf.js and saved as a PNG, which is lossless, so the pixels stay exactly as stored.
| Stored in the PDF as | Saved as |
|---|---|
| JPEG (DCTDecode) | .jpg, the original data unchanged |
| JPEG with a mask or inverted colors | .png, with the mask applied |
| Lossless (Flate, LZW, run length) or uncompressed | .png |
| JPEG 2000, JBIG2 or fax (CCITT) | .png, decoded |
| CMYK, palette or spot colors (not JPEG) | .png, converted to RGB |
Example from a test file built for this tool: four pages with two JPEG photos, three PNG pictures, one inline image and one stencil mask. One photo appears on pages 1 and 2 and one picture on pages 1 and 3, and page 4 holds only a vector rectangle. The tool lists 7 images: both photos as .jpg files identical byte for byte to the photos that went in, the 3 pictures as PNG files with exactly the original pixels (one with its transparency), plus the inline image and the mask. The repeated photo and picture are listed once, each with both of its pages, and the rectangle is not listed, because it is not an image.
How to extract images from a PDF
- 1Drop your PDF on the box or choose it from your device. The pages are read one by one, and each picture appears as it is found.
- 2Check the list: page, pixel size, format and how each image will be saved. Tiny images such as dots and lines are hidden unless you show them.
- 3Press Download under an image to save it, or Download all as ZIP to get every listed image in one file.
To keep a few pages as a smaller PDF instead, use Split PDF. To make a PDF with big photos lighter while keeping its text, use Compress PDF.
Sources
- ISO 32000-1:2008, Document management, Portable Document Format, Adobe copy: 7.4.8 DCTDecode filter, 8.9.5 image dictionaries, 8.9.6.2 stencil masking, 8.9.7 inline images (4 KB or less), 11.6.5.3 soft-mask images
- PDF.js by Mozilla: the PDF engine that reads the pages and decodes the images
Frequently asked questions
Are the images saved in their original quality?
Photos stored as JPEG data in the PDF are saved as .jpg files byte for byte, exactly as they were put into the PDF. All other images are decoded and saved as lossless PNG files with the pixel size stored in the PDF, so nothing is lost on the way. Only an image too large for your browser to draw is saved smaller, and the tool tells you when that happens.
Why is a logo or chart missing from the list?
Logos, charts, icons and diagrams are often drawn as vector graphics: lines, curves and fills, not pixels. They are part of the page, not images, so there is nothing to extract. To get them as a picture, save the page as an image with PDF to JPG and crop it.
Why do I get many strips or tiles instead of one picture?
Some programs split a large picture into strips or tiles when they write the PDF. Each piece is a separate image in the file, so each one is listed on its own. Together they make up the picture you see on the page.
Why does an image have a transparent or odd background?
Images in a PDF can come with a mask. A soft mask holds the transparency, for example the cut-out shape of a product photo, and is kept in the PNG file. A stencil mask is a one-color shape, such as a scanned signature, and is saved as black on a transparent background.
Why is an image listed only once when it is on several pages?
A PDF can store an image once and show it on many pages, like a logo in the header. The tool lists it once and shows all the pages it appears on, so you do not download the same file again and again.
Can I extract the images from a scanned PDF?
Yes. A scanned page is usually one image, so you get one image per page. Its pixel size tells you the scan resolution: an 8.5 inch wide Letter page scanned at 300 dpi is 2,550 pixels wide.
Is my PDF uploaded?
No. Your browser opens the PDF and reads the images from it on your own device. Nothing is sent to a server, so contracts, statements and personal photos stay private.