Remove PDF Background

Remove backgrounds, watermarks and scan specks

1Choose file2Settings3Process4Save

Choose a PDF and its contents will be analyzed and shown here.

Files are processed only in this browser. Nothing is uploaded to a server.

Wipe backgrounds and watermarks by colour or object and clean up scan specks

Continue in another tool
Task settings
Settings
Image objects

Ticked images are removed so the page no longer draws them. Images covering 80% or more of a page (backgrounds, scan watermarks) are ticked in advance.

Preview
Before
After

Remove the password from an encrypted PDF first. Processing focuses on page content, so bookmarks, attachments, forms and signatures may not be preserved. Keep the original and check the result.

About Remove PDF Background

What people call a “background” sits in a PDF in one of two forms. It can be a separate object: a picture laid behind the page, a rectangle filling the whole page, or a diagonal or semi-transparent “SAMPLE” in text. Or it can be part of the same pixels as the text, like the yellowed paper and shadows in a phone scan.

The tool handles those two cases differently. Objects are removed by cutting out the instructions that draw them, so the text stays vector text. Scans are cleaned by re-rendering the page as an image and correcting its colours.

Object removal: what gets cut

The tool reads the page content and cuts only the instructions that draw the images you pick from the list, filled rectangles covering at least 90% of the page, and watermark-like text: text painted with opacity below 1, rotated text, or text matching a phrase or regular expression you give. Every other instruction is copied byte-for-byte, so the body text remains searchable and selectable.

“No longer drawn” is not the same as “gone from the file”, so saving takes one more step. The original content streams from before the cut, and any image no page draws any more, become unreferenced objects, and those are deleted before the file is written. The removed watermark text and background image data therefore don't stay hidden in the file. An image that pages outside the chosen range still draw is, of course, kept. The How PDF Files Work guide explains why a PDF can carry objects that nothing uses.

Scan clean-up: reference colour and tolerance

Each page is rendered (200 dpi by default). The most common colour — usually the paper — is detected, or you pick one yourself, and every pixel whose channels are all within the tolerance (48 by default) of that colour turns white. Specks smaller than a size you set can then be erased and the page brightened. The page is saved as a single JPEG (quality 85 by default) or PNG, so its text can no longer be selected.

Worked examples

A yellowed scan of a contract, tolerance 48

Input
Paper (238,228,200), shadow (205,200,185), pencil signature (170,165,150), blue ballpoint (40,60,160)
Output
Paper and shadow become white; pencil signature and ballpoint stay

The shadow's largest channel difference is 33, under 48, so it goes. The pencil signature's is 68, so it stays. Raise the tolerance to 70 and the signature disappears too — always check the signed page in the preview.

How to use it

  1. Open the PDF to see the page preview and the list of background candidates (images, rectangles, text).
  2. Pick object backgrounds from the list; for scans, set the reference colour, tolerance and speck size.
  3. Check the page range, run it, flip through several pages and save.

Frequently asked questions

My text watermark is not in the list.

It may not be translucent or rotated, or it may be drawn as vector outlines rather than text. Try typing the watermark wording in the phrase field.

The background and the text were scanned as one image.

Then object removal cannot separate them. Use the scan clean-up mode, which works on colour.

What happens to pages with photos?

Light areas of a photo close to the paper colour turn white as well. Leave those pages out of the range.

The file got bigger after scan clean-up.

An A4 page at 200 dpi is about 1,650×2,340 pixels. If the original scan used black-and-white compression (JBIG2 or CCITT), re-saving it as JPEG can make it larger. Try 150 dpi or a lower JPEG quality.

How do I pick the speck size?

Ink blobs smaller than the size you set are erased. Full stops and the dots on i and j are small blobs too, so a large value removes them as well. Start small and raise it while watching the preview.

Last reviewed:

Privacy and how data is handled