The project at-a-glance
- AI-powered app uses GPS to locate lost pets
- Facial recognition tech identifies and matches registered animals
- Designed to alert owners to sightings within set perimeters
- Tech stack: AI (Python), AWS, Flutter, Core 3.0, mysql
- Internal IDS project developed from concept
- Initial roll-out completed in Belarus – global capabilities
When IDS operations director Kseniya Papova had an idea to develop an app to help track down lost pets, she turned the skills gained working within the industry into helping to realise her goal.
Although there are no official published numbers confirming the number of lost cats and dogs each year, estimates by the ASPCA (American Society for the prevention of cruelty to animals) suggest that between 11-16% of dogs and 12-18% of cats are likely to go missing at least once in any five-year period.
A recent analysis published by pet insurers Direct Line also showed that, on average, six dogs are stolen each day across England and Wales — with only 22% returned to their owners.
The requirement of the app was to set-up an international database of owners and pets that would allow for local alerts and searches that narrowed to specific geographical areas and used facial recognition to match found pets to the register — improving the chances of a safe return — all for free.
"What began as a long-term dream of mine to help owners to track down and reunite them with their lost pets, has become a reality across Europe, thanks to the expertise and support of IDS. The realisation of the project has helped to bring AI and facial recognition technology into a new arena.”
"After a successful initial launch in Belarus the app was rolled out worldwide in July 2020. We plan to further develop the technology to improve its accuracy and for Petsi to become the global safe space for pets”
Working with the backing of Belarusian charity Petsi and with the support of the wider IDS management team, Kseniya used the latest AI and facial recognition technology to take the innovative app from concept to market launch.
Users can download the app and register their pets for free, before setting their personalised alerts including geographical areas. Working in a similar way to a traditional lost or found board, when a pet is lost, owners can then use the data stored to search for a match to any reported found pets, view the area where the animal was last sighted and even contact the person who made the post.
Reporting a lost pet is a simple and swift process — a registered cat or dog can be posted in just two clicks.
Likewise, when a lost pet is found, users can use their phones to upload a photo and the GPS information to post an alert on the database. By using facial recognition built into the app, matches on the top ten returned results deliver a 99% accurate match.
Further development and integration are planned, with the goal being the ability to integrate with local animal shelters to assist with the adoption process and a link-up with approved breeders.
in matching pets to database
99% recognition accuracy
within top 10 results