What is the difference between "extract images" and "convert PDF to JPG"?

They are different operations with different outputs. Extract Images pulls each individual embedded image object out of the PDF as its own file at the resolution it was placed in the PDF — original photographs, charts, and logos that the designer embedded. Convert PDF to JPG / PNG rasterises each whole page as a single image at a chosen resolution, baking text and image content together into a page-screenshot. Use Extract Images when you want the original embedded assets back (the photographs, not the page); use page-to-image conversion when you want the page as a viewer sees it.

Will the extracted images be the original resolution or downsampled?

Original resolution — that is the entire point of Extract Images. The tool reads the image streams directly from the PDF without any re-encoding pass; what was embedded comes out unchanged. If the source designer placed a 300 dpi photograph, you get the full 300 dpi back. If they downsampled to 150 dpi before placement, you get 150 dpi back. The PDF's display rendering may have scaled the image to fit a smaller frame on the page, but the underlying image data is preserved at its native resolution.

What formats do the extracted images come out as?

Whatever they were embedded as. JPEGs come out as .jpg with their original quality. PNGs come out as .png with their alpha channel intact. JPEG 2000 images come out as .jp2 (you may need a viewer that handles JP2 — Preview on macOS does; older Windows viewers may not). Bitonal (1-bit) images come out as .png. There is an optional "Convert all to PNG" toggle that re-encodes everything to PNG for compatibility; off by default to preserve fidelity.

Can I extract images from specific pages only?

Yes — the "Pages to scan" field accepts the same range syntax as the rest of the suite: "3", "3-7", "3, 8, 12-15", "1, 30-". The tool walks only the listed pages for embedded image objects. Useful when the PDF has many images and you only want assets from a specific chapter or section.

Is my PDF uploaded to your servers?

No. Extraction runs entirely in your browser using pdf-lib for PDF parsing. Your file is loaded into a sandboxed memory buffer, each page's image objects are enumerated, image streams are decoded and packaged into a zip, and the zip is delivered as a download. Verify in DevTools Network — zero outbound requests. Important for sensitive PDFs whose embedded images may themselves be sensitive (scanned IDs, medical images).

I extracted images but some look low-resolution. Why?

Because the source designer placed a low-resolution image — Extract Images returns what was embedded, not what would have been ideal. Common cause: the PDF was downsampled by an aggressive PDF-compression pass somewhere upstream of you, replacing the original high-resolution images with smaller versions. The original high-resolution images are gone from the file; no extraction tool can recover them. The fix is to go back to the source files (the designer's working files, the original photo library) rather than to the compressed PDF.

How does the tool handle images that are split across multiple PDF objects (tile-based)?

It reassembles them. Some PDF generators (CAD exports, large maps) split a single visual image into tile-objects to manage memory; the tool detects when tile-objects share a parent image and reassembles the full image before extraction. The output is one file per visual image, not one per tile. Without this step, an extracted "tiled" image would arrive as dozens of unhelpful tile fragments — which is what some less-capable extraction tools produce.

Extract Images From a PDF at Full Resolution

7 min read

How to extract images from a PDF (high-quality, original resolution)

By ScoutMyTool Editorial Team · Last updated: 2026-05-20

After working with hundreds of users on image-recovery workflows, the moment "extract the images from this PDF" comes up, the answer usually depends on what kind of "image" you mean. If you want the page as a reader sees it — text, headers, embedded photos all baked into one — you want page-to-image rasterisation (PDF to JPG or PDF to PNG). If you want the original embedded photos back as the standalone files the designer placed in the document, you want true image extraction. The two are different operations with different tools, and below is the workflow for the second one.

Step-by-step: extract the original images from a PDF

The ScoutMyTool tool lives at scoutmytool.com/pdf/pdf-extract-images. Runs client-side — no upload, no signup, no quota.

Drop your PDF. Loads into a sandboxed memory buffer; nothing is uploaded.
Review the image catalogue. The tool walks every page and lists every embedded image object with its native resolution, format (JPEG / PNG / JBIG2 / etc.), and bounding-box size on the page. Thumbnails preview each image.
Pick which images to extract. Default is "all". Select / deselect from the catalogue, or filter by minimum resolution (e.g. "skip anything under 200 px") to drop tiny icons.
Pick output mode. Default: original format preserved (JPEGs as .jpg, PNGs as .png, etc.). Optional: convert all to PNG for downstream compatibility. Optional: convert all to JPEG at chosen quality for size reduction.
Click Extract. The tool reads each selected image's stream from the PDF and writes it to a zip without re-encoding (unless the convert toggle is set). Tile-based images are reassembled automatically.
Download the zip. Each image is named with its page number and order on page (e.g. page-3-img-1.jpg, page-3-img-2.png) for traceability.
Verify a sample at full resolution.Open one of the larger extracted images in an image viewer and zoom in to 100%. The detail should match what the original designer placed — sharp edges, consistent grain, no obvious compression artefacts. If quality is lower than expected, the source PDF was the limit, not the extraction.
If the source PDF is password-protected. Unlock first via Unlock PDF.

Extract Images vs PDF to JPG/PNG — pick the right operation

Extract Images — pulls each embedded image out as a separate file at its original resolution. The output is a set of asset files (the photographs, the charts, the logos as the designer placed them). The output does NOT contain the surrounding text or layout.
PDF to JPG via PDF to JPG — rasterises each whole page as a single JPG at a chosen DPI. The output is page-screenshots with text, images, and headers all baked together into one image per page.
PDF to PNG via PDF to PNG — same as JPG but with transparency preserved. Use when the PDF has transparent regions you want to keep in the output.
The wrong-tool failure mode. Someone asks "extract the images from this PDF" wanting the photographs; the tool they reach for does PDF to JPG; they end up with page-screenshots at low resolution instead of the 300 dpi original photos. Spotting the difference saves a lot of confusion.

What the tool is actually doing (briefly)

The PDF specification (ISO 32000-1) stores embedded images as "Image XObjects" — indirect objects with a stream of pixel data and a header describing dimensions, colour space, and the compression filter applied (DCT for JPEG, Flate for PNG-like, JBIG2 for bitonal scans, etc.)¹. Each page references some subset of these XObjects via the page's resource dictionary and draws them with content-stream operators at the appropriate transformation.

Extraction walks every page's resource dictionary, collects the unique Image XObjects referenced, and writes each one's stream to a file using a header appropriate to its compression filter (DCT → .jpg, Flate → .png after the decompression / reconstruction pass, JBIG2 → .png with bitonal palette). No pixel re-encoding happens unless the user explicitly asks for it.

Frequently asked questions

What is the difference between "extract images" and "convert PDF to JPG"?: They are different operations with different outputs. Extract Images pulls each individual embedded image object out of the PDF as its own file at the resolution it was placed in the PDF — original photographs, charts, and logos that the designer embedded. Convert PDF to JPG / PNG rasterises each whole page as a single image at a chosen resolution, baking text and image content together into a page-screenshot. Use Extract Images when you want the original embedded assets back (the photographs, not the page); use page-to-image conversion when you want the page as a viewer sees it.
Will the extracted images be the original resolution or downsampled?: Original resolution — that is the entire point of Extract Images. The tool reads the image streams directly from the PDF without any re-encoding pass; what was embedded comes out unchanged. If the source designer placed a 300 dpi photograph, you get the full 300 dpi back. If they downsampled to 150 dpi before placement, you get 150 dpi back. The PDF's display rendering may have scaled the image to fit a smaller frame on the page, but the underlying image data is preserved at its native resolution.
What formats do the extracted images come out as?: Whatever they were embedded as. JPEGs come out as .jpg with their original quality. PNGs come out as .png with their alpha channel intact. JPEG 2000 images come out as .jp2 (you may need a viewer that handles JP2 — Preview on macOS does; older Windows viewers may not). Bitonal (1-bit) images come out as .png. There is an optional "Convert all to PNG" toggle that re-encodes everything to PNG for compatibility; off by default to preserve fidelity.
Can I extract images from specific pages only?: Yes — the "Pages to scan" field accepts the same range syntax as the rest of the suite: "3", "3-7", "3, 8, 12-15", "1, 30-". The tool walks only the listed pages for embedded image objects. Useful when the PDF has many images and you only want assets from a specific chapter or section.
Is my PDF uploaded to your servers?: No. Extraction runs entirely in your browser using pdf-lib for PDF parsing. Your file is loaded into a sandboxed memory buffer, each page's image objects are enumerated, image streams are decoded and packaged into a zip, and the zip is delivered as a download. Verify in DevTools Network — zero outbound requests. Important for sensitive PDFs whose embedded images may themselves be sensitive (scanned IDs, medical images).
I extracted images but some look low-resolution. Why?: Because the source designer placed a low-resolution image — Extract Images returns what was embedded, not what would have been ideal. Common cause: the PDF was downsampled by an aggressive PDF-compression pass somewhere upstream of you, replacing the original high-resolution images with smaller versions. The original high-resolution images are gone from the file; no extraction tool can recover them. The fix is to go back to the source files (the designer's working files, the original photo library) rather than to the compressed PDF.
How does the tool handle images that are split across multiple PDF objects (tile-based)?: It reassembles them. Some PDF generators (CAD exports, large maps) split a single visual image into tile-objects to manage memory; the tool detects when tile-objects share a parent image and reassembles the full image before extraction. The output is one file per visual image, not one per tile. Without this step, an extracted "tiled" image would arrive as dozens of unhelpful tile fragments — which is what some less-capable extraction tools produce.

Extract the original images from your PDF now — free, no signup, no upload

Original-resolution, original-format, no re-encoding. Runs entirely in your browser.

Open the Extract Images tool at scoutmytool.com/pdf/pdf-extract-images →