We’ve just perceived how AI image-processing algorithms can be imbecilic and keen simultaneously, and that is on full show with NVIDIA’s most recent interactive demo. “GANimal” gives you a chance to snap a photo of your preferred pet and move its expression to any irregular creature or breed including pugs, hyenas, bears and lions. That animal will then “smile” or posture simply like your Fluffy or Fido.
“Most GAN-based image translation networks are trained to solve a single task. For example, translate horses to zebras,” said NVIDIA’s lead computer-vision researcher Ming-Yu Liu in a blog. “In this case, we train a network to jointly solve many translation tasks, where each task is about translating a random source animal to a random target animal. Eventually, the network learns to generalize to translate known animals to previously unseen animals.”
On the off chance that you’ve seen AI picture interpretation demos previously, you may figure that the outcomes won’t be uniform. In numerous cases, it puts your pet’s charming articulation accurately on different creatures and in others, not really. With NVIDIA’s own model (over) different mutts and even the hyena look about right, yet the sloth bear turned out a piece, contorted and the wild bear has a yellow tongue.
In any case, it shows guarantee as an approach to show AI how to manage obscure subjects and be more improvisational. The team next plans to use it on other kinds of images, like flowers and food, at higher resolutions. “This is how we make progress in technology and society by solving new kinds of problems,” said Liu. Hopefully, that won’t include the creation of freaky hybrid animals — check here to try it for yourself.