Why document photos always have shadows
A flatbed scanner uses a controlled internal light source that illuminates the page uniformly. When you photograph a document with a phone or camera, you're working with ambient light — and that almost never falls evenly.
- Your hand or phone casting a shadow while holding the device over the page
- A single-direction light source — desk lamp, window — creating a gradient across the page
- Thick documents or books where pages curve away from flat, shadowing the center crease
- Glossy or laminated documents reflecting light unevenly across the surface
The result: one corner is bright, the other is dark. The text is readable in places, but the overall image looks nothing like a proper scan.
How the shadow removal works
PhotoToScan uses a machine learning model trained specifically on document images — not general photography. This matters because document shadows behave differently from shadows in regular photos.
- Document boundary detection — separating the paper from the background first
- Lighting gradient analysis — understanding which parts are in shadow vs. well-lit
- Paper color inference — determining what the white background should look like under uniform light
- Text preservation — correcting the surrounding lighting while keeping ink sharpness intact
This is not a brightness or contrast adjustment. Those are global operations — they make everything brighter uniformly, which blows out the already-lit areas. PhotoToScan applies spatially-aware correction: shadowed regions are brightened while lit regions are left unchanged.
How to use it
- Step 1 — Upload your document photo (JPG, PNG, HEIC, WebP)
- Step 2 — PhotoToScan automatically detects and removes the shadow
- Step 3 — Preview the result, then download as JPG or PDF
No account required to preview. Works in any browser, on any device.
When to use this
- Government forms you need to submit digitally
- Contracts and legal documents being shared via email
- Academic transcripts or certificates uploaded to application portals
- ID cards and passports where shadow can cause digital submission rejection
- Receipts and invoices for expense reporting
- Pages from books or binders where the binding creates center shadows
Frequently asked questions
- Does it work on photos taken in poor lighting?
- Yes. The model handles both strong directional shadows and general underexposure. If your document is readable by eye, PhotoToScan can produce a clean result.
- Will it affect the text quality?
- Shadow removal is applied to the background lighting layer, not the text. Text sharpness is preserved — in most cases, contrast actually improves because the dark shadow is no longer competing with the ink.
- Is this the same as adjusting brightness and contrast?
- No. Brightness/contrast are global — they make everything brighter or darker uniformly. PhotoToScan applies spatially-aware correction, so shadowed regions brighten without blowing out areas that were already well-lit.
- Can it handle shadows in multiple areas at once?
- Yes. The model processes the entire document in a single pass and corrects all affected regions simultaneously.
- What file formats are supported?
- JPG, PNG, HEIC, and WebP as input. Output can be downloaded as PDF or JPG.