How to Resize Images Inside a PDF (And When You Should)

Images are usually the biggest contributor to a PDF's file size, and they are frequently the wrong size, too big for distribution, too small for print, or oversized at one resolution while another image on the same page is undersized. Resizing images in a PDF is a workflow most people only need occasionally, but when they need it, the difference between doing it well and doing it badly is dramatic.

This guide covers the practical paths for resizing, both visual resizing on a page and underlying pixel-resolution resampling, and the trade-offs between them.

Two different "resize image" operations

When people say they want to resize an image in a PDF, they usually mean one of:

Visual resize. Shrink or grow the image's footprint on the page. The pixel data is unchanged; the rendering scales it up or down.
Resolution resize (resample). Reduce the underlying pixel data. A 4000×3000 photo becomes a 1000×750 photo, saving disk space and download time.

Both reduce visual size on a page; only resolution resize reduces the file size.

For shrinking the file overall, see reduce PDF file size. For just adjusting layout, this article is the right place.

Visual resize: tools

Adobe Acrobat Pro. Tools → Edit PDF → click the image → drag the corner handles to resize. Hold Shift to preserve aspect ratio. The pixel data stays the same; only the display dimensions change.

Foxit PDF Editor. Edit → Edit Object → click the image → use the resize handles.

LibreOffice Draw. Open the PDF, click the image, drag handles. Re-export as PDF.

Browser-based. Docento.app supports clicking images and resizing them in the browser.

After visual resize, the file usually stays the same size on disk (the image bytes are still there). Sometimes it grows slightly because the editor rewrites the content stream.

Resolution resize: tools

To actually shrink the embedded image data:

Adobe Acrobat Pro. File → Save As Other → Reduced Size PDF (one-click downsample), or File → Save As Other → Optimized PDF for control over per-image settings.

Ghostscript. A typical command to downsample all images:

gs -dPDFSETTINGS=/ebook -sDEVICE=pdfwrite \
   -o smaller.pdf input.pdf

/ebook downsamples to 150 DPI; /printer to 300 DPI; /prepress keeps images near original quality. See our Ghostscript introduction.

pdfsizeopt, open-source PDF optimizer that focuses heavily on image resizing.

Custom scripts. Using pikepdf or PDFKit, extract each image, resize with Pillow or ImageMagick, and re-embed. Best for precise control over specific images.

When to use which

Image looks too big on the page. Visual resize. The pixel data is fine; you just want a smaller footprint.
PDF is too large to email. Resolution resize. Drop image DPI to 150 or 96.
Photo will be printed at full quality. Leave alone. Resizing down would lose quality.
Photo was overdone, exported at 600 DPI for a screen-only PDF. Resolution resize to 150 DPI.
Image is for thumbnails or icons. Resolution resize aggressively to 96 DPI or lower.

Print vs screen resolution

Two reference points for image DPI:

300 DPI, print standard. Sharp output at typical viewing distance. Often the right target for documents that will be printed.
150 DPI, common compromise for documents read both on screen and occasionally printed. Visually acceptable, file-size friendly.
96 DPI, screen-only. Tiny files, fine for on-screen reading, mediocre for print.
72 DPI, the historical "web" DPI. Suitable for low-quality web display, rarely a deliberate choice today.

A typical 2400×1800 photo embedded in a PDF:

At 300 DPI display size, occupies 8 × 6 inches on the page
At 150 DPI display size, occupies 16 × 12 inches
The choice between the two is which size makes sense at your target page dimensions and viewing distance

Resizing a single image precisely

Acrobat Pro lets you click an image and enter exact dimensions in the Properties panel. Tools → Edit PDF → click image → Properties → enter width/height in inches, cm, or pixels. Constrain proportions to avoid stretching.

For programmatic precision:

import pikepdf
from PIL import Image
import io

with pikepdf.open("input.pdf", allow_overwriting_input=True) as pdf:
    for page in pdf.pages:
        for name, img_obj in page.images.items():
            pdf_image = pikepdf.PdfImage(img_obj)
            pil_image = pdf_image.as_pil_image()
            # resize to 1000 px width, preserving aspect
            ratio = 1000 / pil_image.width
            new_size = (1000, int(pil_image.height * ratio))
            resized = pil_image.resize(new_size, Image.LANCZOS)
            buf = io.BytesIO()
            resized.save(buf, format='JPEG', quality=85)
            img_obj.write(buf.getvalue(), filter=pikepdf.Name('/DCTDecode'))
    pdf.save("output.pdf")

Resizes every image to a maximum of 1000 pixels wide. Adjust as needed for your use case.

What to preserve when resizing

Each image carries metadata that may matter:

Color profile, ICC profile that defines how to interpret the colors. Preserve it during resize to avoid color shifts.
EXIF data, camera metadata for photos. Usually safe to strip, but check first if provenance matters.
Transparency, PNG alpha channel. Preserve for images with transparency.
Mask, image mask defining transparent areas. Preserve along with transparency.

Tools differ on what they preserve by default. For brand-critical color work, verify the output by comparing the colors on a test page.

Common gotchas

Visual resize without re-embedding. Most editors update the image's display size in the content stream without changing the image data. File size barely changes. To shrink the file, you need resolution resize.

Different DPI per image. A single PDF may have images at 600, 300, and 96 DPI mixed together. Downsampling globally to 150 DPI affects all of them, sometimes shrinking the 96 DPI ones unnecessarily. Use a tool that respects an "if larger than" threshold.

Text-on-image confusion. A scanned page is one big image of text. Resizing the image of text directly affects readability. Lower DPI = blurrier text. For scanned pages, 200-300 DPI is the floor; below that, text becomes hard to read.

Vector graphics treated as images. Some PDF tools flatten vectors into raster during resize. The resulting "image" can be resized further, but you have lost the resolution-independence of the original. Check whether your tool preserves vectors.

Aspect ratio distortion. Without explicitly constraining proportions, dragging a corner handle can stretch the image. Always hold Shift (or use the appropriate modifier in your tool) for proportional resize.

Layered overlays. A PDF page may have multiple images stacked (e.g., a chart over a background image). Resizing one without the others creates visible gaps or overlaps. Inspect the layering before resizing.

Resizing images in scanned PDFs

A scanned PDF where each page is a full-page image, the most common output of a scanner, has different needs:

Downsample with caution. Drop from 600 DPI to 300 DPI for typical office documents. Below 300, text legibility suffers.
JPEG quality. Re-encoding at 75-85% quality is typically imperceptible and significantly smaller.
Color depth. A bilevel (black and white) scan compresses far better than grayscale or color. If your scanned text is essentially black on white, convert to bilevel during resize. See how to convert a PDF to TIFF for related concepts.

For OCR concerns, see PDF OCR explained, downsampling a scanned PDF before OCR can hurt accuracy.

Replacing an image entirely

Sometimes the cleanest "resize" is to replace the image. Take the image at its target resolution and swap it in. See how to replace an image in PDF.

Quick recipes

Shrink a 50 MB image-heavy PDF to email size:

gs -dPDFSETTINGS=/ebook -sDEVICE=pdfwrite -o small.pdf input.pdf

Resize a single image visually without changing pixels:

Open in Acrobat Pro
Tools → Edit PDF → click the image
Drag corner handles with Shift held
Save

Globally cap all images at 150 DPI:

Use Acrobat's Optimized PDF (Image Settings) or a Ghostscript variant. Set color/grayscale/bitonal DPI targets.

Takeaway

Resizing images in PDFs splits into visual resize (cheap, fast, no file-size change) and resolution resize (the real win when file size matters). Acrobat Pro and Foxit handle visual resize well; Ghostscript and pikepdf are the right tools for resolution resize. Always preserve aspect ratio. For scanned PDFs, downsample with care so text stays legible. After resizing, verify both the visual layout and the file size, and check that any signatures you might want later have not been invalidated. For browser-based image resizing alongside other edits, Docento.app handles the operation in one place.