From Alex J. Champandard: As seen on TV! What if you could increase the resolution of your photos using technology from CSI laboratories? Thanks to deep learning and #NeuralEnhance, it's now possible to train a neural network to zoom in to your images at 2x or even 4x. You'll get even better results by increasing the number of neurons or training with a dataset similar to your low resolution image.
The catch? The neural network is hallucinating details based on its training from example images. It's not reconstructing your photo exactly as it would have been if it was HD. That's only possible in Hollywood — but using deep learning as "Creative AI" works and it is just as cool! (github)
Zoya Teirstein for The Verge: The Supreme Administrative Court of Sweden just ruled that camera drones qualify as surveillance cameras and require a permit under Sweden's camera surveillance laws. The ruling requires owners to cough up a sizable fee in order to get their equipment off the ground, and paying to start the process is no guarantee a citizen will be granted the right to fly. County administrators will have to consider whether use of a "surveillance camera" overrides the public's right to privacy on a case-by-case basis.
Aerial photographers and recreational drone users may have to rely on alternative methods to get their footage. The justices said dash cams and cameras affixed to bicycle handlebars are not in violation of the public's right to privacy because the devices are within reaching distance of the people who operate them.
The ruling targets recreational and commercial users alike, and makes zero exceptions for journalists. Sweden's leading drone company Unmanned Aerial System (UAS) said up to 3,000 people may lose their jobs as a result of the court's decision. Cont'd...
Intel CEO Brian Krzanich to Speak, The Weather Channel's Jim Cantore to Host AUVSI's XPONENTIAL 2017
"Future of Drones: Clearing the Flight Path to a Connected World" to Be Opening Day Keynote at Drone World Expo
From James Charles, Derek Magee, David Hogg:
The objective of this work is to build virtual talking avatars of characters fully automatically from TV shows. From this unconstrained data, we show how to capture a character’s style of speech, visual appearance and language in an effort to construct an interactive avatar of the person and effectively immortalize them in a computational model. We make three contributions (i) a complete framework for producing a generative model of the audiovisual and language of characters from TV shows; (ii) a novel method for aligning transcripts to video using the audio; and (iii) a fast audio segmentation system for silencing nonspoken audio from TV shows. Our framework is demonstrated using all 236 episodes from the TV series Friends (≈ 97hrs of video) and shown to generate novel sentences as well as character specific speech and video... (full paper)
David Hambling for Popular Mechanics: Unmanned Warrior is the world's biggest robot war game, currently taking place for two weeks off the coast of Scotland. It was proposed by First Sea Lord Admiral George Zambellas to give airborne, surface, and underwater drones from various suppliers a chance to show off their prowess. Unmanned Warrior is part of Joint Warrior, an exercise involving 30 warships and submarines from 18 nations. But for the newly inaugurated robot portion, the U.S. is a strong presence, with teams from the Office of Naval Research (ONR).
As robots continue packing increased capability in smaller, lower-cost packages, these exercises are a glimpse into the future of naval warfare. Of course, none of these drones are armed for these exercises. These Unmanned Warriors are not trusted to carry weapons—at least not yet.
The first-ever Unmanned Warrior will wrap up on October 20. Cont'd...
Records 61 to 75 of 2475