Skip to content

OCR Text Scanning

Learn how to use optical character recognition (OCR) to extract text, dates, and other information from your photos.

Overview

The OCR scanning feature uses AWS Rekognition to detect and extract text from images. This is particularly useful for:

  • Handwritten dates on the back of photos
  • Photo studio stamps and watermarks
  • Captions or notes written on prints
  • Location names or event labels

Scanning Images

Automatic Scanning on Upload

When you upload an image with an optional back image, OCR scanning runs automatically:

  1. Upload front image - Your main photo
  2. Back image auto-detected - System searches for matching back image files
  3. Combined scan - Both images are scanned together
  4. Results displayed - Text and dates shown in a dialog

Loading Indicator

A loading spinner appears during OCR processing. Wait for it to complete before reviewing results.

Back Image Detection

The system automatically searches for back images using common naming patterns:

Front Image Back Image Patterns Searched
photo.jpg photo_back.jpg, photo-back.jpg, photo_b.jpg, photoback.jpg
IMG_001.jpg IMG_001_back.jpg, IMG_001-back.jpg, IMG_001_b.jpg

Supported patterns:

  • _back suffix (e.g., image_back.jpg)
  • -back suffix (e.g., image-back.jpg)
  • _b suffix (e.g., image_b.jpg)
  • back suffix without separator (e.g., imageback.jpg)

Manual Back Image Selection

If the system doesn't find your back image automatically:

  1. Click the Browse button next to "Back Image"
  2. Select your back image file
  3. The system will scan both images when you upload

OCR Results Dialog

After scanning, a results dialog appears showing:

Detected Text

All text found in the image(s), including:

  • Handwritten notes
  • Printed text
  • Stamps and labels

Deduplication

Duplicate text entries are automatically removed when combining front and back scans.

Detected Dates

Dates extracted from the text with:

  • Formatted date - Standardized date format
  • Confidence score - How certain the detection is (percentage)
  • Date type - Classification of the date format

Using the Results

Apply to Metadata:

  • Click Use Date to apply a detected date to the image metadata
  • The most reliable date is automatically suggested first
  • Edit the description field to include relevant text

Dialog Controls:

  • Draggable header - Click and drag the header bar to move the dialog
  • View image underneath - Reposition dialog to see your photo while reviewing

Date Parsing

Supported Date Formats

The OCR system recognizes many date formats:

Format Type Examples Priority
Full named date Sep 12 1964, September 12, 1964 Highest
ISO format 1964-09-12 Highest
Month/Day/Year 09/12/1964, 9-12-1964 High
Month and Year Sep 1964, September 1964 High
Holiday dates Christmas 1964, Easter 1965 Medium
Year only 1964 Medium
Short year formats Sep '64, 9/12/64 Lower

Date Priority

When multiple dates are detected, they are sorted by reliability:

  1. Named dates (e.g., "Sep 12 1964") - Most reliable
  2. ISO dates (e.g., "1964-09-12") - Most reliable
  3. Month/Year (e.g., "Sep 1964") - High reliability
  4. MDY format (e.g., "09/12/1964") - High reliability
  5. Holiday dates - Medium reliability
  6. Year only - Medium reliability
  7. 2-digit year formats - Lower reliability (ambiguous century)

4-Digit vs 2-Digit Years

The system prioritizes 4-digit years to avoid ambiguity:

  • "Sep 12 1964" → Parsed as 1964 ✅
  • "Sep 12 '64" → Parsed as 1964 (assumes 1900s for dates before 30)
  • "Sep 12 64" → May be ambiguous without context

2-Digit Year Ambiguity

Photos with clear 4-digit years will get the highest confidence scores. When possible, ensure dates in your photos include the full year.

Draggable Dialogs

Both the OCR Results dialog and Metadata Editor are draggable:

How to Move Dialogs

  1. Click the header bar - The dark header at the top of the dialog
  2. Drag to new position - Move the dialog anywhere on screen
  3. Release to place - Dialog stays in the new position

Why Draggable?

  • View the image - Move the dialog to see your photo while reviewing OCR text
  • Compare results - Position dialogs to compare detected text with the image
  • Comfortable editing - Place dialogs where they're most convenient

Tips for Best Results

Image Quality

  • Clear, high-resolution images work best
  • Good lighting improves text detection
  • Minimal blur helps accuracy

Text Characteristics

  • Printed text is detected more reliably than handwriting
  • Dark ink on light background works best
  • Larger text is more accurately recognized

Back Image Scanning

  • Photograph the back of prints to capture dates and notes
  • Use consistent naming (e.g., photo.jpg and photo_back.jpg)
  • Include the back image during upload for automatic combined scanning