A scanned PDF is just a pile of page images — there is no real text inside, so it cannot reflow or be searched. To turn it into a proper EPUB, you need OCR (optical character recognition) to read the words out of the images first. Here is how the process works and what to expect.
Why Scanned PDFs Are Hard
A normal PDF stores text as characters; a scanned PDF stores each page as a photo of the page. Without OCR, a converter sees pictures, not words, so the EPUB you get is just images — no selectable text, no reflow, no working table of contents. OCR is the step that makes the text real.
What OCR Does
OCR scans each page image and guesses the characters, producing a text layer behind the picture (or replacing the picture with text). Once the words exist as text, the converter can build paragraphs, detect headings, and create a navigable EPUB. Accuracy depends on scan quality — crisp, straight scans OCR far better than skewed, low-resolution ones.
Step-by-Step: Scanned PDF to EPUB
- Check the source — if the PDF already has selectable text (try copying a word), skip OCR and convert directly with PDF to EPUB.
- Run OCR — use an OCR tool or service to add a text layer to the PDF.
- Convert to EPUB — feed the OCR'd PDF to PDF to EPUB; a formatting-aware path keeps images and layout where possible.
- Proofread — scan the output for OCR errors (common with unusual fonts or footnotes) and fix in the source if needed.
The PDF to EPUB keep-formatting guide explains the settings that preserve layout.
When OCR Isn't Enough
OCR is not magic. Heavily illustrated pages, two-column layouts, and footnotes often come through messy, and the converter cannot always tell where a chapter starts. For a clean result:
- Pre-process the scan — deskew and raise the resolution before OCR.
- Fix headings in the source so the EPUB TOC builds correctly.
- Accept some cleanup — complex books may need a manual pass.
If the conversion still misbehaves, the common ebook format problems guide maps each symptom to a fix.
Key Takeaways
- A scanned PDF is images, not text — it needs OCR before it becomes a real EPUB.
- OCR adds a text layer the converter can turn into paragraphs and a TOC.
- Scan quality decides accuracy — crisp, straight, high-res scans OCR best.
- Use a formatting-aware path — see PDF to EPUB keep-formatting.
- Proofread the output — OCR errs on odd fonts and footnotes.