For Mohit Arora, Global Digital Marketing Lead at Pet Nutrition, what’s worse than losing a pet is the lack of adequate infrastructure to find lost pets and bring them back to their families. While dogs are synonymous with loyalty and being man’s best friend, a lack of speech is a deficiency that causes many pet dogs to disappear every year.
While a collar with emergency details is an easy solution, customized collars are not very popular yet. A lot of dogs reject anything more than a simple light-weight collar around their necks. And again, there is always the risk of sabotaging your missing dog collar. However, with the advancement of technology, especially machine learning, losing your dog too soon can be a problem we never have to deal with.
That’s exactly what the brilliant minds behind ForPAWS are working on. ForPAWS is a facial recognition tool by Mars Incorporated that will help reunite lost pets with their families. Ahead of the app’s wider rollout, Arora explained the basics of how facial recognition will soon become a reality for canines, as well as why the need for such devices exists today.
Requirement of ForPAWS
Forpaws, as defined by Aurora, is “a free service mobile app developed by Mars Pet Nutrition that leverages facial recognition technology and a network of pet-lovers to help track missing pets and bring them home.” to help solve a serious problem.”
“With ForPAWS we make sure that concerned pet parents always have the most reliable help available to find their pets and bring them back if they ever go missing. Initial offers are for dogs only. We are working on a module for cats that will be rolled out in the future,” explains Arora.
“With ForPAWS you only need a photo and a basic smartphone to create a unique ID for your dog,” says Arora.
How will an app let you find your missing dog?
ForPAWS uses machine learning (an artificial intelligence-based implementation that lets software tools accurately interpret situations and predict outcomes) to scan a pet’s facial details from a large database. This database will be filled with unique pet IDs created by registering their facial data, making the device equivalent to a school/college/office identity card.
However, this identity card will be one of several in a digital avatar that the app’s servers can access on demand. So, should your dog go missing, anyone who finds him will be able to use ForPAWS to trace his furry friend back to you, after which he can contact you directly.
“Our data scientists trained machine learning (ML) modules for breed recognition and facial recognition. In addition to making predictions more accurate, they use indicative markers for a particular dog such as color, spots, etc. ML is currently trained on more than 130 breeds and 2,000 unique dogs,” explains Arora, “As the number of pet images in our database grows, ML must become smarter.”
Eliminates the process of going through tons of listings to find out whose missing pet you’ve met.
“One challenge we face and where we have made a recommendation to our community of pet owners is to take a new photo of your dog every 6 to 12 months and match the ML to the most recent one. Allowing the picture to be used, as dogs grow faster than humans.
A question of effectiveness, and of secrecy!
A third party research conducted on more than 750 dogs, including “dogs of the breed, stray and mix of indies,” returned a 95% match at a rate of over 90 percent. “To further improve the accuracy, we are looking at indicative markers for a particular dog in terms of color, spots and the location where the dog was lost,” says Arora.
Coming to how the service’s privacy mechanics will work, Arora explains that the app will be clear about how it works and will implement and maintain an ‘opt in and opt out’ approach. “We feel the privacy of consumers and giving them the option to “opt in and opt out” on data sharing is a reciprocal way of delivering a free digital service,” says Arora.
“If the consumer changes his mind after using the application, we have an opt out option in the application where the consumer can ask us to remove a particular, or all, information based on his comfort level. This way We give the consumer the power to decide how much they want to share with us.”
While pet parents find a useful ‘lost and found’ platform, we want to be able to use data to improve pet and pet-parent experiences in both the digital and offline worlds in our endeavor , which they are comfortable sharing with us. Create a better world for pets.
Machine learning is half the job; What about user adoption?
For services like ForPAWS to work, a large database and a customized machine learning algorithm by itself may not be enough. The tool will also rely heavily on user adoption, and will work best when the platform has a large community and network of pet lovers registered.
To make sure ForPAWS gets there quickly, the company is taking several additional steps, in addition to launching a free app. These include focusing on targeted geography at the city-level and ensuring that there is a large enough community that “can act as a searcher in case a pet is lost.”
The tool is also getting rescues, shelters and vets to register on the app, as these are usually the first places searchers take a missing pet to seek help. Arora explains, “We already have the largest coverage of NGOs and rescues out of Bengaluru – CUPA, CARE, Cubbon Park Canine, who have supported the effort and are looking to distribute the app among their volunteers. “
“To support a pet parent who has lost a pet, ForPAWS provides a virtual poster that parents can push to their social handles (Facebook, WhatsApp, Twitter, etc.). In addition, we are also helping them expand their reach to find pets by using the social handles of Mars PetCare and its partner communities on Facebook that help find lost dogs.
ForPAWS is currently available on Android, and while the app’s functionality is still limited to the city of Bengaluru, it is expected to reach more cities and even countries other than India in the near future.