Making Add Me a actuality required collaboration between the Pixel Digital camera, Inventive Digital camera and Google XR groups. (The Google XR crew works on Android XR and ARCore, platforms for constructing augmented and digital actuality experiences.) To mix a picture, Add Me makes use of augmented actuality (AR) to indicate the second photographer an overlay of the primary picture to allow them to precisely body the brand new photograph to match the composition of the primary one. Since aligning the 2 photographs is essential to get the very best consequence, the crew explored a number of strategies to get the right shot. They finally realized augmented actuality was essentially the most helpful option to visually information the consumer, so that they turned to the Google XR crew.
“We on the XR crew have at all times wished to do extra and allow useful options within the Digital camera app,” says XR software program engineer Ryan DuToit. The crew noticed the imaginative and prescient for Add Me and acknowledged how helpful it might be for private images.
Nonetheless, creating an interface the place the AR function was self-explanatory — so even these unfamiliar with the know-how may use it — wasn’t straightforward, and took ample experimentation. Plus, the XR crew had a busy 2024, Ryan explains, pointing to the latest Android XR launch. “Collaborating with one other crew on one other challenge was difficult — however we did it, and I’m very pleased with it,” he says.
Add Me was a reasonably difficult technical feat: It makes use of numerous machine studying fashions, that are powered by Pixel’s TPU and the Tensor G4 chip to run on system. “If it weren’t for the TPU, I don’t assume we’d have been capable of converge to an affordable latency,” Adi says. In the event that they’d needed to run these ML fashions on a GPU or CPU, the function wouldn’t have been capable of show the pictures from the primary shot after which later the blended shot shortly sufficient.
