Special formats

How to combine text and image clues in a source search

Readable text can lead to a source even when the image is too generic for visual matching. Search distinctive words separately, then compare the resulting page with the photograph.

Transcribe conservatively

Record only text you can actually read. Mark uncertain characters instead of inventing them. A caption, asset code or unusual phrase may be more useful than a common word such as “sale”.

Combine the trails

  1. Search the full image.
  2. Search a distinctive quoted phrase where appropriate.
  3. Open candidate publications.
  4. Compare the image, text and surrounding context together.

A poster with a unique event title may be easy to locate through text even if its background photograph appears on hundreds of pages. The event page can then supply the correct credit.

Avoid circular confirmation

Several sites may repeat the same incorrect caption. Repetition is not independent verification. Look for the original publisher or a primary record, and keep the text match separate from the visual match until both support the same source relationship.

Continue with finding an original image source.

Published by Facehunch with AI assistance. How we use sources and maintain these guides.
KEEP EXPLORING

Related guides

FROM READING TO EXPLORING

See how a photo search comes together.

Try the fictional example, or run a consented API test. Paid searches are disabled.

Open Facehunch