BRENNEN

Industries / Retail, hospitality and personal services / Pet care

Tracking and identification

Giving each pet a permanent number and a registered owner, carrying that identity across borders, and finding a pet that has gone missing.

7

Milestones achieved, from 1996 to 2025.

3

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1996

    One standard for chips

    ISO 11784:1996 and ISO 11785:1996 set how an animal microchip, a transponder injected under the skin, encodes its number and answers a reader at 134.2 kHz. European pet travel rules accepted chips built to them from 2004, and England's chipping laws require them.

    legislation.gov.ukeur-lex.europa.eu

    Credit

    International Organization for Standardization (ISO)

  2. 2004

    A passport for pets

    From 3 July 2004, dogs, cats and ferrets traveling with their owners between European Union countries needed an identification mark and a passport issued by an authorized vet showing a valid rabies vaccination. A readable tattoo was accepted for an eight-year transition, after which a microchip became the only accepted mark.

    eur-lex.europa.eu

    Credit

    European Parliament and Council, Regulation (EC) No 998/2003

  3. 2008

    Scanners miss some chips

    US shelters met chips at three frequencies: 125, 128 and 134.2 kHz. Researchers chipped and scanned 3,949 dogs and cats at six shelters. No scanner found every chip of any brand. The universal scanners found at least 94.8% of 128 and 134.2 kHz chips, but one found only 66.4% to 75.0% of 125 kHz chips.

    Lessonrival chip formats meant a chip could be present and still not read, and the authors advised scanning each animal more than once.

    ebi.ac.uk

    Credit

    Linda K. Lord and colleagues (Journal of the American Veterinary Medical Association, 2008)

  4. 2009

    Chips bring strays home

    In 53 US shelters between August 2007 and March 2008, owners were found for 74.1% of 2,956 microchipped stray dogs and 63.5% of 469 stray cats. Where owners were not found, the main reasons were a wrong or disconnected phone number (35.4%), no reply (24.3%) and a chip registered to another group (17.2%). Of the chips that shelters looked up with a registry, only 58.1% were registered at all.

    Lessona chip is only as useful as the owner record behind it.

    ebi.ac.uk

    Credit

    Linda K. Lord and colleagues (Journal of the American Veterinary Medical Association, 2009)

  5. 2016

    England requires dog chips

    From 6 April 2016, every keeper of a dog older than 8 weeks in England had to have it microchipped with an ISO-standard chip and its details recorded on a database.

    legislation.gov.uk

    Credit

    UK Parliament, the Microchipping of Dogs (England) Regulations 2015

  6. 2024

    England adds cats

    From 10 June 2024, keepers of cats older than 20 weeks in England must have them microchipped under rules that replaced the dog regulations and cover both species.

    legislation.gov.uk

    Credit

    UK Parliament, the Microchipping of Cats and Dogs (England) Regulations 2023

  7. 2025

    A tracker brand switched off

    Whistle, a maker of GPS pet trackers owned by Mars, was bought by the tracker company Tractive, and every Whistle tracker stopped working on 31 August 2025. Owners could ask for a free Tractive replacement until 30 September 2025.

    Lessona GPS collar works only while its maker runs the network behind it, so a sale or shutdown can end it overnight.

    engadget.com

    Credit

    Whistle, Mars and Tractive; reported by Engadget

Watch for

  1. Chips for every EU pet

    Regulation (EU) 2026/1818 requires every dog and cat kept in the European Union, including pets, to carry a microchip and be registered in a national database, and the national databases must work with each other. The Commission must also build a free public system that lets a buyer check an advertised animal's chip and registration. The duties start on 31 August 2030 for breeders and sellers, 31 August 2036 for dog owners and 31 August 2041 for cat owners.

    Achieved whenthe verification system is live and member states' databases answer each other's lookups, due from 31 August 2031.

    eur-lex.europa.eu

    Claimed by

    European Parliament and Council of the European Union, June 2026

  2. Photo matching finds pets

    Petco Love says Petco Love Lost, a free database that matches photos of lost pets against photos of strays reported by finders and shelters, has reunited more than 200,000 pets with owners since 2021. It counts a reunion only when the person who posted the pet comes back and clicks a button to report it, and it told the American Animal Hospital Association that it believes only about a third of reunions get confirmed that way.

    Achieved whenan outside group checks a sample of the claimed reunions against shelter records or publishes how often the system's matches are right.

    dailynews.co.zaaaha.org

    Claimed by

    Petco Love, March 2026, as reported by The Washington Post

  3. Doorbell cameras look for dogs

    Ring, Amazon's doorbell camera brand, says its Search Party feature, which checks nearby owners' outdoor cameras for a reported lost dog, has helped reunite more than one dog a day since it launched, and in February 2026 opened it to anyone in the US.

    Achieved whenRing publishes how many searches it ran, alerts it sent, false alerts and confirmed reunions, or an outside audit does.

    aboutamazon.com

    Claimed by

    Amazon, February 2026

Open

  1. 01

    Pair photos of reunited pets

    A public set of real reunion photo pairs: the photo an owner had on their phone and the photo a shelter took at intake, of the same lost dog or cat, collected with the owner's consent when they reclaim the pet. Mix the pairs with thousands of other intake photos and score open matching models on how often they rank the right animal first and in the top ten.

    The brief
    Credit · unclaimed
  2. 02

    Bench-test GPS pet trackers

    A public accuracy test of the trackers sold for pets. A person carries several trackers on a dummy collar alongside a survey-grade satellite receiver along fixed routes in open country, suburbs, a city center and indoors, and compares each tracker's reported positions with the true track. No animal is needed.

    The brief
    Credit · unclaimed

Open 01 · the brief

Pair photos of reunited pets

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public set of real reunion photo pairs: the photo an owner had on their phone and the photo a shelter took at intake, of the same lost dog or cat, collected with the owner's consent when they reclaim the pet. Mix the pairs with thousands of other intake photos and score open matching models on how often they rank the right animal first and in the top ten.

Why it is still open

Research sets and models exist, such as PetFace (2024, faces of 257,484 animals), LostPaw (2023, 90% accuracy on its own dog images) and a 2026 study reporting 84.28% top-1 accuracy on a 1.9-million-photo corpus. Each reports accuracy on its own research images. Kashtanka.pet, a volunteer project that matches lost and found ads, built its test pairs by splitting the photos of a single ad, because it could not know which ads showed the same animal. The two largest services, Petco Love Lost and Ring Search Party, publish reunion counts but not how often their matches are right.

arxiv.orgarxiv.orgarxiv.orgmn.uio.no

Built with what exists

the PetFace benchmark code; the LostPaw model (github.com/vandrw/lostpaw-transformer); open image encoders such as SigLIP 2; fine-tuning a small open vision model on one GPU (see the Note "One GPU to fine-tune"); Hugging Face to host the set; a one-page consent form.

  1. First step

    Ask one county shelter that photographs every stray at intake to collect 100 owner photos with consent at reclaim.

  2. Done looks like

    1,000 consented reunion pairs, 20,000 distractor intake photos, top-1 and top-10 scores for at least five open models, and the set released under an open license.

Open 02 · the brief

Bench-test GPS pet trackers

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public accuracy test of the trackers sold for pets. A person carries several trackers on a dummy collar alongside a survey-grade satellite receiver along fixed routes in open country, suburbs, a city center and indoors, and compares each tracker's reported positions with the true track. No animal is needed.

Why it is still open

The one peer-reviewed accuracy test found tested low-cost CatLog data loggers for wildlife research, which recorded positions for later download, and found a mean location error of 15.4 m. For the live trackers owners buy, such as Tractive, Fi, Garmin or Apple AirTag, a search found only review sites. One anonymous site describes testing eight trackers against a survey-grade receiver but publishes no raw tracks, and other review sites rank the same trackers differently.

ebi.ac.ukdog-gps-tracker-reviews-com.pages.dev

Built with what exists

an RTK receiver board such as the u-blox ZED-F9P, which fixes position to within a few centimeters, as the reference; the trackers and their apps, with location history exported or screen-logged; Python with pandas and pyproj to compute errors.

  1. First step

    Walk one 5 km loop with three trackers and a phone logging raw satellite data, and compare the tracks.

  2. Done looks like

    at least five trackers, 10 km of route in each of four settings, the median and 95th-percentile position error per device, the share of updates more than five minutes old, and the raw tracks published.