Conventional methods for pet identification, such as tags and microchips, have their flaws. Tags can easily come loose, and not all pet owners are comfortable with the idea of microchipping their pets. Even those who are comfortable may encounter issues like chip damage and outdated databases.
In response to this challenge, Jesse Joonho Lim and Ken Daehyun Pak introduced an app called Petnow, designed to identify cats and dogs by scanning their facial features. Petnow, which has secured $5.25 million in funding from Daedeok Venture Partners and DigiCap at a valuation of $24 million, is currently participating in Startup Battlefield 200 at TC Disrupt 2023.
Before co-founding Petnow in 2018, Lim was a co-leader in a semiconductor startup named Chips&Media, which later pursued an IPO. Pak, who also holds a doctorate in electrical engineering like Lim, had a career spanning over a decade as an AI video processing researcher before joining Petnow.
At its core, Petnow operates by conducting a camera-based facial scan of a pet through a mobile app available on Android and iOS devices. Utilizing AI trained on approximately 200,000 images of dog and cat faces, collected both by the Petnow team and sourced from users’ pets, the app creates a unique biometric profile for each pet.
To address concerns about accidentally capturing other objects or individuals in the background, Petnow claims to employ an algorithm that automatically detects and focuses on dogs or cats while excluding any extraneous elements.
For dogs, Petnow records a “nose print,” asserting that a dog’s nose is as distinctive as a human fingerprint and remains unchanged over time, providing a reliable means of distinguishing between individual dogs. In the case of cats, Petnow examines a cat’s “facial contour,” which the company contends retains its distinctiveness due to cats’ unique grooming habits (though this claim may appear less certain than the concept of a dog’s nose print).
Lim and Pak envision Petnow being used for tasks such as pet registration without the need for a veterinarian, locating lost pets, and generating “pet IDs” to verify insurance status.
“The pet identification market may not be mature at the moment, but it will eventually become big,” Lim and Pak stated in an email interview with TechCrunch. “Pet identification technology is a fundamental product that can be continuously used by people, unlike products or services that go viral only for short periods of time. The market has great potential because pets should have IDs like people, and their data can be backed up to form an ultimate pet platform.”
Petnow claims that its algorithms boast a “99% accuracy” rate in identifying individual cats and dogs. However, it’s well-documented that even the most advanced image-analyzing AI systems can exhibit bias, whether intentional or not.
For instance, facial recognition technology has led to the mistaken arrests of at least six Black individuals due to biases introduced during training, often stemming from inadequate representation of Black faces in datasets or the use of mugshot databases featuring a disproportionate number of Black faces photographed in poor lighting conditions, hindering the algorithm’s ability to differentiate between individuals.
Setting aside the specific challenges of facial recognition, even experts struggle to differentiate between breeds of animals, let alone animals of the same breed. A recent study involving 5,000 dog experts nationwide found that only a small minority could correctly identify one of the breeds by analyzing their DNA.
Petnow asserts that its training database continues to expand, and it employs AI to ensure that pet photos are captured with optimal brightness and sharpness. The company also emphasizes its commitment to data privacy, stating that it does not share a user’s or pet’s information with third parties without the user’s consent and offers an option to delete stored data at any time. Petnow points to a study co-authored by its data scientists in the journal IEEE Access, which indicates that its dog nose-print identification technology achieved over 99% accuracy in distinguishing between noses. However, this study dates back to 2021 when the training dataset was presumably smaller, and the research on its cat facial recognition algorithm has yet to be made public.
While Petnow has garnered approximately 70,000 users, its adoption among pet care providers and shelters has been relatively slow, with only five undisclosed enterprise and public sector customers in France, Toronto, and South Korea, where the company is based. Petnow is currently in a pre-revenue stage, with a monthly burn rate of $150,000. However, it anticipates securing a contract with Korean domestic and international pet insurers by October, as well as launching pilots in France and collaborating with a metropolitan government in Canada for their pet registries.
Lim and Pak believe that there is substantial room for growth in the pet identification market, especially considering the rapid increase in pet ownership during the pandemic. They anticipate that government pet registry programs and partnerships with pet insurance providers will contribute to the expansion of their product across North America and Europe.
In conclusion, the effectiveness of Petnow’s technology remains a critical question. Petnow asserts its algorithms are highly accurate, but the challenges of bias and the difficulty of distinguishing between similar animals present potential concerns. Given the potential consequences, it is crucial that Petnow and similar technologies deliver on their promises of accurate pet identification to avoid misleading pet owners, which would be a particularly cruel form of deceptive advertising.
