Whether you’ve photographed a whiteboard after a meeting, received a scanned business card, or need to extract text from a screenshot, converting images to editable text is one of the most frequently needed productivity tasks. The good news is that the tools for doing this have improved dramatically — your smartphone, Google Drive, and numerous online services can all perform OCR with surprising accuracy. This guide covers every major method, with practical tips for maximizing accuracy and handling special cases like tables and mathematical formulas.
Method 1: Google Lens (Best Quick Mobile Option)
Google Lens is the most frictionless way to extract text from an image on a smartphone. It’s available on all Android phones (often accessible directly from the camera), and on iOS through the Google app or Google Photos.
How to use it:
- Open Google Lens — from the camera viewfinder, from Google Photos by tapping an image and selecting the Lens icon, or from the Google app search bar.
- Point at text or open an existing photo.
- Select the “Text” option. Google Lens highlights all detected text.
- Tap “Select all” then “Copy” to copy everything, or select specific portions of text.
- Paste into any app — notes, email, a document, or a messaging app.
Google Lens also offers a live camera mode for real-time text recognition — useful for menus, street signs, or anything you want to read or translate instantly without saving a photo first. Accuracy on printed text is excellent; handwriting recognition has improved significantly but remains less reliable than printed text recognition.
Method 2: Microsoft Lens (Best for Document Scanning)
Microsoft Lens (formerly Office Lens) is more specialized than Google Lens — it’s designed specifically for scanning documents and whiteboards, with features tailored to professional document handling.
How to use it:
- Download Microsoft Lens from the App Store or Google Play (free on both platforms).
- Choose the capture mode: Document (for paper documents), Whiteboard (enhanced contrast for white backgrounds), or Business Card (extracts contact information automatically).
- Photograph the document. Microsoft Lens auto-detects the page edges and applies perspective correction, straightening even photos taken at an angle.
- Choose your save format: Word document (editable, with OCR applied), PDF (with searchable text layer), or OneNote page.
The perspective correction feature is particularly valuable. If you photograph a book or document from an angle — rather than directly overhead — Microsoft Lens flattens the image to a frontal view, significantly improving OCR accuracy on content that would otherwise be distorted.
Method 3: Online OCR Tools (Best for Desktop Use)
For processing images on a computer, web-based OCR tools are fast and require no software installation:
- iLovePDF / iLoveIMG: Upload JPG, PNG, or other image formats and receive searchable PDF or plain text output. Supports multiple languages.
- OnlineOCR.net: Free tool supporting 46 languages, with output in Word, Excel, or plain text format.
- Google Drive: Upload an image (JPG or PNG) to Drive and open it with Google Docs — Drive automatically applies OCR and makes the text editable.
- Adobe Acrobat online: Convert images to searchable PDFs with high accuracy, though an Adobe account is required.
Accuracy Optimization: Image Quality is Everything
OCR accuracy correlates directly with input image quality. Here’s how to systematically maximize the quality of your source images:
Resolution and Sharpness
- For scanner-produced images, use at least 300 DPI. For documents with small text (under 10pt), 400-600 DPI produces meaningfully better results.
- When photographing with a smartphone, ensure the camera is in focus before shooting. Move closer to the document rather than zooming in digitally — optical proximity produces sharper results than digital zoom.
- Use your phone’s dedicated document scanning mode (available in iOS Notes, Files by Google, and many camera apps) rather than the regular photo mode. Document scanning modes apply automatic sharpening and edge detection optimized for text.
Contrast and Lighting
- High contrast between text and background is critical. Light gray text on white paper dramatically reduces accuracy.
- Avoid harsh side lighting or glare, especially when photographing glossy magazines or laminated documents. Even, diffuse frontal lighting produces the cleanest images.
- If your source image has poor contrast, increase contrast and sharpness using a photo editing app (Snapseed, Adobe Lightroom Mobile, or even your phone’s built-in editor) before running OCR.
Alignment and Cropping
- Tilted text is one of the most common causes of OCR errors. When scanning, ensure the document lies flat. When photographing, use the on-screen grid lines to align the document’s edges with the frame.
- Crop out empty margins and non-text areas before processing. Reducing visual noise helps the OCR engine’s layout analysis step work more accurately.
Table Recognition: Preserving Structure
Tables are one of the hardest challenges for general-purpose OCR tools. Most tools extract the text content but lose the row-and-column structure, producing a meaningless stream of numbers and labels instead of a usable table.
- Adobe Acrobat Pro: Reasonably good at preserving table structure when exporting to Word or Excel.
- ABBYY FineReader: Table recognition is a core strength — it produces well-structured Excel output from tabular images more reliably than most alternatives.
- Microsoft Azure Form Recognizer (Document Intelligence): An AI-powered cloud API specifically designed for structured forms and tables. Far more accurate than general-purpose OCR for tabular data, and worth the cost for high-volume table extraction workflows.
Mathematical Formula Recognition
Standard OCR tools cannot reliably recognize mathematical formulas — the combination of Greek letters, operators, subscripts, superscripts, and fractions is too complex for character-by-character recognition. Specialized tools exist for this purpose:
- Mathpix Snip: Screenshot or photograph a mathematical expression and Mathpix converts it to LaTeX code or MathML with impressive accuracy. It has become a standard tool for students in mathematics, physics, and engineering.
- Microsoft OneNote: Recognizes handwritten mathematical expressions and converts them to formatted equations, useful for converting handwritten notes to typed documents.
- Wolfram Alpha: Can solve equations photographed with a smartphone, with built-in formula image recognition as part of its interface.
Do it in your browser — the file never leaves your device: PDF Tools
