UncChapelHill
Latest
AI can predict autism through babies' brain scans
Scientists know that the first signs of autism can appear in early childhood, but reliably predicting that at very young ages is difficult. A behavior questionnaire is a crapshoot at 12 months. However, artificial intelligence might just be the key to making an accurate call. University of North Carolina researchers have developed a deep learning algorithm that can predict autism in babies with a relatively high 81 percent accuracy and 88 percent sensitivity. The team trained the algorithm to recognize early hints of autism by feeding it brain scans and asking it to watch for three common factors: the brain's surface area, its volume and the child's gender (as boys are more likely to have autism). In tests, the AI could spot the telltale increase in surface area as early as 6 months, and a matching increase in volume as soon as 12 months -- it wasn't a surprise that most of these babies were formally diagnosed with autism at 2 years old.
iSpy software can read texts and steal passwords with its little eye (video)
We spy, with our bleary eye, a new piece of software that could make it dramatically easier to steal personal data. The program, known as iSpy, allows devious voyeurs to remotely identify and read text typed on touchscreen displays. That, in and of itself, isn't exactly new, but iSpy takes shoulder surfing to slightly terrifying new areas -- namely, those beyond the "shoulder." Developed by Jan-Michael Frahm and Fabian Monrose of the UNC-Chapel Hill, this program, like those before it, takes advantage of the magnified keys found on most touchscreens. All you'd have to do is point a camera at someone else's screen and iSpy will automatically record whatever he or she types by stabilizing the video footage and identifying the enlarged keys. If you're using a smartphone camera, you'll be able to eavesdrop from up to three meters away, but if you opt for a more heavy duty DSLR device, you could steal passwords from up to 60 meters away. The software can also recognize any words typed into a device, and, according to its architects, can identify letters with greater than 90 percent accuracy. When used with a DSLR camera, iSpy can even pick up on reflections of touchscreens in sunglasses or window panes from up to 12 meters away. To avoid this, Frahm and Monrose recommend disabling the magnified key function on your smartphone, or using some sort of screen shield. We recommend checking out a video of the program, after the break.
UNC researchers develop a system for creating 3D models using images pulled from Flickr, off-the-shelf components
A group of researchers from the University of North Carolina and the Swiss university ETH-Zurich have teamed up to develop a system for creating 3D models of famous landmarks using photos from photo sharing websites like Flickr. Unlike previous projects at Microsoft and the University of Washington, the team at UNC used a home PC (albeit one with four GPUs) to process millions of images pulled from the Internet and construct 3D models of such landmarks as the Colosseum and the Roman baths at Sagalassos (above). And all the models were created in less than a day. According to UNC Chapel Hill's Jan-Michael Frahm, the process improves on current commercial systems by a factor of 1,000 to one. "Our technique would be the equivalent of processing a stack of photos as high as the 828-meter Dubai Towers, using a single PC, versus the next best technique, which is the equivalent of processing a stack of photos 42 meters tall – as high as the ceiling of Notre Dame – using 62 PCs," he said. "This efficiency is essential if one is to fully utilize the billions of user-provided images continuously being uploaded to the Internet." He sees any number of uses for this technology, from AR integration to 3D maps for rescuers in case of a natural disaster.
