How We Removed Spam from “Find a Friend”
In the “Find a Friend” project, we introduced deduplication based on visual similarity to merge identical sightings, reduce noise, remove duplicates, and avoid flooding the owner with unnecessary notifications.
In real life, it looks like this: the same found cat gets photographed by 5 passersby, and the owner receives 5 urgent notifications. Not very demure 😅
We ran into this in the “Find a Friend” project and introduced deduplication based on visual similarity. The system compares photos and understands that shots taken by different people belong to the same animal.
What this gave us:
- automatic merging of listings
- less information noise
- fewer duplicates in search
- higher trust in the platform
In short, we built anti-spam not for emails, but for cats. Against the backdrop of the boom in AI features in products, this is exactly the kind of case where CV solves a real user pain point, not something “for hype.”
Result: the owner gets one normal notification instead of five clones — and finds their furry friend faster.