Why Image Quality Matters
OCR accuracy is fundamentally limited by the quality of the input image. Even the best engine cannot reliably read text in a blurry, dark, or skewed photo. Here are five concrete improvements you can make today.
1. Use the Highest Resolution Available
More pixels mean more detail. When photographing a document, get close enough to fill the frame — don't photograph a document on a table from two metres away. If scanning, use at least 300 DPI (600 DPI for small text or handwriting).
Rule of thumb: Individual characters should be at least 20–30 pixels tall in the final image.
2. Get the Lighting Right
- Use natural daylight (not direct sunlight which creates glare)
- Avoid flash — it creates hotspots
- If indoors, position the document near a window
3. Keep the Document Flat
Curved pages — from a book or a thick folder — distort text baselines and reduce accuracy. Press the document flat, or use a glass surface to hold it down. Most modern OCR tools handle mild skew, but strong curves cause problems.
4. Maximise Contrast
Dark text on a white background is easiest to read. If your document is a photo of text on a coloured surface, try increasing contrast in any image editor before uploading. Many phone cameras have a "document" or "whiteboard" scan mode that does this automatically.
5. Crop Out Distractions
The more of the image that contains background (table, floor, hands), the lower the signal-to-noise ratio. Crop your image tightly around the text before uploading. This also speeds up processing.
Checking Your Results
TextLens returns a confidence score with each extraction. If you see a score below 85%, try re-photographing with better lighting before concluding the tool has limitations.