|Sign up now!|
Kiwi programmers have successfully taught a computer to distinguish between a cat’s face and its rear-end in a move developers say will help improve relationships with the notoriously aloof animal.
The team of programmers who created the digital technology spent months training the computer to make the physical distinction.
The three strong team leveraged an artificial intelligence network which can identify objects within images. The developers labelled over 8,000 images showing dozens of different breeds of cats from different angles. They then trained a deep neural network by manually highlighting the unique characteristics of the cat’s features within each image until software could make the distinction by itself.
This field of machine learning technology also has applications in protecting wildlife - for example a computer can be taught to identify a poacher stalking a tiger from a real time video feed streamed from drones - using facial recognition software to identify the animal and a human face in close proximity. The software can then automatically alert authorities with the location - giving them the chance to rescue the animals.
Lead developer David Arcus said while there was a light hearted purpose behind this version of the technology, they hope that others will be inspired to use it for the benefit of animals.
“What we wanted to do was find a fun way to bring cat owners closer to their pet.
“Kiwi cat owners can upload an image of their cat to a free website set up by the developers. If the computer’s AI identifies it as the rear end of a cat, it determines that the owner needs a closer relationship with their pet and will send them a new type of food product called Creamy Treats designed to be hand-fed to cats; thereby giving them more face-to-face time with their pet.
Arcus says while cats can have a reputation for being antisocial, some experts suggest this is a myth. The reason behind this is a natural protection mechanism, where they are always alert to their surroundings and remain constantly onguard - something humans misinterpret as disinterest.