Duke University researchers have developed an AI tool that can turn blurry, unrecognizable pictures of people’s faces into eerily convincing computer-generated portraits, in finer detail than ever before.
Previous methods can scale an image of a face up to eight times its original resolution. But the Duke team has come up with a way to take a handful of pixels and create realistic-looking faces with up to 64 times the resolution, ‘imagining’ features such as fine lines, eyelashes and stubble that weren’t there in the first place.
“Never have super-resolution images been created at this resolution before with this much detail,” said Duke computer scientist Cynthia Rudin, who led the team.
The system cannot be used to identify people, the researchers say: It won’t turn an out-of-focus, unrecognizable photo from a security camera into a crystal clear image of a real person. Rather, it is capable of generating new faces that don’t exist, but look plausibly real.
While the researchers focused on faces as a proof of concept, the same technique could in theory take low-res shots of almost anything and create sharp, realistic-looking pictures, with applications ranging from medicine and microscopy to astronomy and satellite imagery, said co-author Sachit Menon ’20, who just graduated from Duke with a double-major in mathematics and computer science.
The researchers will present their method, called PULSE, at the 2020 Conference on Computer Vision and Pattern Recognition (CVPR), held virtually from June 14 to June 19.
Traditional approaches take a low-resolution image and ‘guess’ what extra pixels are needed by trying to get them to match, on average, with corresponding pixels in high-resolution images the computer has seen before. As a result of this averaging, textured areas in hair and skin that might not line up perfectly from one pixel to the next end up looking fuzzy and indistinct.
The Duke team came up with a different approach. Instead of taking a low-resolution image and slowly adding new detail, the system scours AI-generated examples of high-resolution faces, searching for ones that look as much as possible like the input image when shrunk down to the same size.
The team used a tool in machine learning called a “generative adversarial network,” or GAN, which are two neural networks trained on the same data set of photos. One network comes up with AI-created human faces that mimic the ones it was trained on, while the other takes this output and decides if it is convincing enough to be mistaken for the real thing. The first network gets better and better with experience, until the second network can’t tell the difference.
PULSE can create realistic-looking images from noisy, poor-quality input that other methods can’t, Rudin said. From a single blurred image of a face it can spit out any number of uncannily lifelike possibilities, each of which looks subtly like a different person.
Even given pixelated photos where the eyes and mouth are barely recognizable, “our algorithm still manages to do something with it, which is something that traditional approaches can’t do,” said co-author Alex Damian ’20, a Duke math major.
The system can convert a 16×16-pixel image of a face to 1024 x 1024 pixels in a few seconds, adding more than a million pixels, akin to HD resolution. Details such as pores, wrinkles, and wisps of hair that are imperceptible in the low-res photos become crisp and clear in the computer-generated versions.
The researchers asked 40 people to rate 1,440 images generated via PULSE and five other scaling methods on a scale of one to five, and PULSE did the best, scoring almost as high as high-quality photos of actual people.
See the results and upload images for yourself at http://pulse.cs.duke.edu/.
Sony Corp. presented Thursday the world’s first image sensors with built-in artificial intelligence, promising to make data-gathering tasks much faster and more secure.
Calling it the first of its kind, Sony said the technology would give “intelligent vision” to cameras for retail and industrial applications.
The new sensors are akin to tiny self-contained computers, incorporating a logic processor and memory. They’re capable of image recognition without generating any images, allowing them to do AI tasks like identifying, analyzing or counting objects without offloading any information to a separate chip. Sony said the method provides increased privacy while also making it possible to do near-instant analysis and object tracking.
Sony joins technology giants like Huawei Technologies Co. and Alphabet Inc.’s Google in working to build dedicated AI silicon to help accelerate everything from image processing to machine learning. Its latest semiconductors could offer a big boost to augmented reality applications, should the technology be adapted for the smartphone or consumer markets, where the Japanese company is a leader.
The new AI-augmented sensors are capable of capturing a regular 12-megapixel image, 4K video at up to 60 frames per second or neither, providing only metadata about what the sensor has seen. Among the applications suggested by Sony are the counting and tracking of visitors to public spaces, heat and congestion mapping and measuring shopper behavior in retail locations.
Though intended for commercial customers in its present iteration, the technology has promise for consumer applications as well. It could help a personal device such as a smartphone identify objects and users securely, without generating any actual images. The accelerated object detection would also be an advantage for maintaining sharp focus when filming fast-moving subjects such as sports players or pets.
Sony is the world leader in providing image sensors for smartphones such as Apple Inc.’s iPhone and dedicated photo and video cameras from the likes of Nikon Corp.
Its sensor division has been its most reliable growth driver over the past few years, boosted by the proliferation of multicamera phones.
The new products are in line with the company’s long-term goal, as articulated by Chief Executive Officer Kenichiro Yoshida, of expanding the variety of sensing solutions it offers and pursuing more forms of recurring revenue.
Sony said it had already shipped samples of its new sensors to potential customers, who are mostly in the business-to-business segment including factory automation.
More information: PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models, arXiv:2003.03808 [cs.CV] arxiv.org/abs/2003.03808
