Extract Images From PDF

How the Extract Images From PDF works

A PDF does not store a page as one picture. Text, vector drawings and images are separate objects, and each photo or scan inside the file is kept as its own image stream with a width, a height, a colour space and a compression filter. Extracting those streams gets you the pictures at the exact resolution they were placed in the document, which is usually far sharper than a screenshot.

How to extract images from PDF files here

The tool opens the PDF with pdf-lib and walks every object in the file, not just the ones on visible pages, and picks out each image stream. What happens next depends on how the image was compressed:

JPEGs are copied byte for byte, so there is no second round of compression and no quality loss. Other images are decoded to raw pixels, converted from their colour space to RGB and written as PNG. Grey, RGB, CMYK, ICC-based and indexed (palette) images are all handled, at 1, 2, 4, 8 or 16 bits per channel. If an image has a soft mask, the mask becomes the PNG's transparency, so a logo with a transparent background keeps it.

One image comes back as a single file. Two or more are packed into a zip named after the PDF, with files numbered in the order they appear in the file's object table.

Real uses

• Recovering photos from a report or brochure when the original image files are lost.
• Reusing a logo or chart that was supplied only inside a PDF.
• Checking image resolution before printing: the extracted pixel size tells you how much detail is really there, regardless of how large it is drawn on the page.
• Pulling scanned pages out of a scanner's PDF as plain image files for editing.
• Auditing what a file contains. Hidden or off-page images are found too, which matters before you share a document that was edited from someone else's template.

Tips and pitfalls

Vector graphics are not images. Charts made in Excel, diagrams from design tools and most text are drawn with vector instructions, so they are not extracted. To capture those, take a screenshot of the page at a high zoom level instead.

Small images are often decoration. Bullets, icons, rules and tiny gradient tiles are images too. The default setting skips anything under 50 pixels on its shorter side; choose Keep every image if you need them.

Duplicates are stored once. A logo repeated on every page is usually one object in the PDF, so it is extracted once, which is what you want.

CMYK conversion is approximate. Print-ready PDFs often hold CMYK images. They are converted to RGB with a simple formula rather than a colour profile, so strong colours may look slightly different from the printed result. JPEG CMYK images are copied as is and keep their original data.

Masked and stencil images. Images used only as transparency masks are not saved on their own, and one-bit stencil masks, which a PDF fills with a colour, are skipped because they hold no colour of their own.

Fax and JBIG2 scans. Some scanners store black-and-white pages with CCITT or JBIG2 compression. Browsers cannot decode these, so they are counted in the note instead of being saved.

How to read the result

The stats show how many images the PDF holds, how many were saved, how many were original JPEGs and how many were written as PNG, how many were skipped for being small, and the download size. A note lists any images in formats that could not be converted. If the count is lower than you expect, the missing pictures are probably vector drawings, not images.

Related tools