Google Pixel's cameras: “The generative AI will allow us to solve problems that could never have been before”
The subjects at the center of Google's concerns, at least on the telephony, are varied: between its artificial intelligence Geminiits rise in power as a manufacturer of equipment and its work in the world of shooting – the three subjects being intimately linked -, Mountain View is on all fronts. Her Pixel 9a freshly formalized crystallizes his work, and if he is not the real flagship of the brand – it is more affordable and logically less perfected than Pixel 9 Pro XLfor example -, it testifies to the rapid advance in the work of the American firm.
Accelerate the pace
As Isaac Reynolds confirms to us, Lead Product Manager From the Pixel Camera branch within Google, with whom we were able to maintain ourselves, the pace accelerated at the edge of the 2020s-that is to say more or less when Google began to display international ambitions for its Pixel smartphones. “In the past two years, we have crossed a turning point, thanks to the access we have to researchers' teams [R&D] from Google ”he specifies.
This narrow collaboration seems to be the engine of innovation. “We have so close relationships with researchers' teams that our global way of working, prioritizing, building and making decisions is optimized to use elements from research articles published six months earlier and transform them into functions on our products”. It is therefore this proximity that allows Google to quickly integrate the latest discoveries of technical research in its smartphones. However, it is impossible to know how many people make up these famous teams of researchers who allow the firm to move forward by step as a giant: this is a manufacturing secret.
AI at the heart of the photographic “revolution”
And if research is advancing so fast, it is because AI is invited to the party. Reynolds cites the example of the Auto Frame function (automatic framing), which places the subject at the center of the image and extends if necessary its frame by filling the space with automatically generated content. “These are things that we could never have done with computational imaging”welcomes Isaac Reynolds. “The generative AI will allow us to solve problems that we could never have previously sent, such as the photos in low light or the digital zoom”he specifies.
What will be the new technologies to come? “I am so curious to know what they will be”admits Isaac Reynolds, stressing that even in Google, the future of photography is constantly exploring. “The beauty of research is that you don't know what it can bring before you have it before you”.
While the multi-frame (the aggregation of several images) constituted the strength of the first pixel, the latest born of Google rely on artificial intelligence. “We can use AI to reduce noise [en basse lumière]and use it to improve the zoom produces more true details than with what I call traditional techniques ”. From the next hardware (larger sensors, new optics, etc.) or software on the horizon, we will not know anything more, except that the semantic analysis of the images will remain at the heart of the work of the American giant. To better create products “Facts to capture memories, and not only to capture light passing through a photosensitive surface”enthuses the manager of the Pixel Cameras section, a bit philosopher. It remains to be seen if the Pixel 9a – when available in the trade– will be up to these creative ambitions.
