Petnow: Revolutionizing Pet Identification with AI-Driven Technology
Petnow, a startup, claims to identify pets by scanning their faces, addressing the limitations of conventional pet identification methods.
The app has raised $5.25 million in funding and is a participant in the Startup Battlefield 200 at TC Disrupt 2023.
It works by creating a biometric profile using AI-trained on 200,000 pet snout images.
The technology, while promising, raises concerns about accuracy and potential biases.
Petnow emphasizes data privacy and user consent.
The startup aims to expand globally and enter partnerships with pet insurance providers.
Main AI News:
In the realm of pet identification, the conventional methods have shown their flaws. Tags get misplaced, and not everyone embraces the idea of microchipping their beloved companions. Even when microchipping is chosen, issues like chip malfunction and outdated databases can pose significant challenges. However, two innovators, Jesse Joonho Lim and Ken Daehyun Pak, have stepped forward to introduce a game-changing solution through their startup, Petnow. This groundbreaking app, which has garnered $5.25 million in funding from Daedeok Venture Partners and DigiCap, boasts a valuation of $24 million and is currently making waves as a participant in the Startup Battlefield 200 at TC Disrupt 2023.
Before embarking on their journey with Petnow in 2018, Lim had co-led a semiconductor startup, Chips&Media, which later achieved an IPO. Meanwhile, Pak, also holding a doctorate in electrical engineering like Lim, had devoted over a decade to the field of AI video processing research before joining the Petnow team.
In essence, Petnow operates by utilizing a camera-based scan of a pet’s face via a mobile app, compatible with Android and iOS devices. Leveraging the power of artificial intelligence, Petnow has been trained on a vast collection of approximately 200,000 images of dog and cat snouts, sourced from both the Petnow team and user submissions. This data enables Petnow to create a unique biometric profile for each pet.
To alleviate concerns about accidental captures, such as a family member standing behind the pet, Petnow employs an advanced algorithm that automatically identifies and focuses on dogs or cats while excluding any extraneous background.
For dogs, Petnow records what they call a “nose print.” Remarkably, they assert that a dog’s nose is as distinctive as a human fingerprint and remains unchanged over time, establishing it as a dependable means of distinguishing between different canines. When it comes to cats, Petnow examines their “facial contour,” asserting that this feature maintains its distinctiveness due to the unique grooming habits of each feline. (Though, some might find this claim about cats a tad more questionable than the reliability of a dog’s nose print.)
Lim and Pak envision Petnow being employed for various purposes, including registering pets without requiring a visit to the vet, locating lost pets, and creating “pet IDs” to verify insurance status. They emphasize that while the pet identification market may currently be in its nascent stage, its potential for growth is substantial. Pet identification technology is a fundamental and enduring product that can serve people continuously, unlike fleeting viral products or services. In their view, pets should have IDs similar to humans, and this data can form the foundation of the ultimate pet platform.
Yet, the critical question remains: does Petnow’s technology live up to its claims? Petnow boldly asserts that its algorithms achieve an impressive “99% accuracy” in identifying individual cats and dogs. However, it is well-documented that even the most sophisticated image-analyzing AI systems are susceptible to bias, whether intentional or inadvertent.
As a poignant example, facial recognition technology has led to the mistaken arrests of at least six individuals, all of whom were Black. These misidentifications stem from the inadequate representation of Black faces in the training datasets, introducing biases. Similarly, in the animal kingdom, even experts grapple with distinguishing between breeds, let alone animals of the same breed. A recent study involving 5,000 dog experts nationwide revealed that only a small minority could correctly identify the breeds through DNA analysis.
To address these concerns, Petnow asserts that its training database continually expands, and it employs AI to ensure that pet photos are captured under optimal lighting conditions for enhanced accuracy. Petnow also emphasizes its commitment to user data privacy, affirming that it does not share user or pet information with third parties without consent and offering the option to delete stored data.
Petnow does reference a study published in the journal IEEE Access, co-authored by its data scientists, which attests to the accuracy of its dog nose-print identification technology, claiming over 99% precision in distinguishing between canine noses. However, this study dates back to 2021 when the training dataset was presumably smaller, and the research for its cat face-recognizing algorithm remains unreleased.
The stakes are undeniably high. An algorithmic error could obstruct a family’s quest to find a missing pet or lead a veterinarian to access the wrong animal’s vaccination records. While Petnow has garnered approximately 70,000 users to date, it has secured only five undisclosed enterprise and public sector clients in France, Toronto, and its home base in South Korea.
Petnow, presently in its pre-revenue stage with a monthly expenditure of $150,000, anticipates signing a contract with domestic and international pet insurers in Korea by October. Additionally, it plans to launch pilot programs in France and collaborate with a metropolitan government in Canada for its pet registries.
Petnow’s innovative approach to pet identification offers exciting potential for the market. However, concerns about accuracy and biases must be addressed to gain widespread acceptance. With its expansion plans and partnerships, Petnow is positioning itself to play a significant role in shaping the future of pet identification technology. The evolving pet identification market may witness transformative changes with the rise of such groundbreaking solutions.