Photo Library Search
Typing 'dog', 'beach' or 'receipt' into your photo app and finding matching pictures you never tagged, powered by image classifiers and image embeddings.
Also Google Photos · image search · image embeddings · find dogs · photo tagging
Google Photos (2015); similar features in Apple Photos and others
Your phone looks at every photo and quietly notes what is in it, so later you can just ask for 'birthday cake' and it finds them.
Where it fits
+All relationships as a list
Is a kind of / part of
Siblings
Uses / built on
How it works
When Google launched Google Photos in May 2015, a headline feature was search without tagging: type “dog” and see your dog photos, because a deep neural network had looked at every picture. Apple Photos and other apps offer the same today.
Two technologies work together:
- Image classification and object detection assign labels like dog, beach, food to each photo.
- Image embeddings go further. Models like CLIP place photos and text in one shared embedding space, so a query like “red car in the snow” can find matching photos even without a “red car in the snow” label. Searching that space is a vector search problem.
The feature also shows the stakes of errors. In 2015, Google Photos labeled photos of Black people as “gorillas”. Google apologized; the reported fix was to remove the label rather than solve the underlying problem, a well-known case study in bias and fairness.
Everyday use cases
- Case 01
How you've probably met it
Searching your phone's gallery for "dog" or "snow" or "passport" and getting results without ever labeling a photo.
Strengths, weaknesses and when to use it
Strengths
- Makes huge personal photo libraries searchable with zero effort.
Weaknesses
- Mislabels can be hurtful; in 2015 Google Photos notoriously labeled Black people as gorillas, and Google's fix was to block the label.
When to use itAs the everyday example of image classification and embeddings at consumer scale.
Compare with
Go deeper
Sources
Facts checked as of 2026-10-12