Fast, Accurate Detection of 100,000 Object Classes on a Single Machine a prizewinning paper by Google Research scientists, describes a breakthrough in machine vision that can distinguish between a huge class of objects 20,000 times faster than before.
This so-called convolution operator is one of the key operations used in computer vision and, more broadly, all of signal processing. Unfortunately, it is computationally expensive and hence researchers use it sparingly or employ exotic SIMD hardware like GPUs and FPGAs to mitigate the computational cost. We turn things on their head by showing how one can use fast table lookup — a method called hashing — to trade time for space, replacing the computationally-expensive inner loop of the convolution operator — a sequence of multiplications and additions — required for performing millions of convolutions with a single table lookup.
We demonstrate the advantages of our approach by scaling object detection from the current state of the art involving several hundred or at most a few thousand of object categories to 100,000 categories requiring what would amount to more than a million convolutions. Moreover, our demonstration was carried out on a single commodity computer requiring only a few seconds for each image. The basic technology is used in several pieces of Google infrastructure and can be applied to problems outside of computer vision such as auditory signal processing.
Fast, Accurate Detection of 100,000 Object Classes on a Single Machine
(Image: Clutter, a Creative Commons Attribution Share-Alike (2.0) image from neofob's photostream)
Gabriella Coleman is the “hacker anthropologist” whose book on the anthropology of Anonymous is among the best books on hacking I’ve ever read; her new paper in Current Anthropology, From Internet Farming to Weapons of the Geek, poses a fascinating question: given that hackers are as well-paid and privileged as doctors, lawyers and academics, how […]
In SoundNet: Learning Sound Representations from Unlabeled Video, researchers from MIT’s computer science department describe their success in using software image-recognition to automate sound recognition: once software can use video analysis to decide what’s going on in a clip, it can then use that understanding to label the sounds in the clip, and thus accumulate […]
The University of Toronto’s Citizen Lab (previously) is one of the world’s leading research centers for cybersecurity analysis, and they are the first port of call for many civil society groups when they are targeted by governments and cyber-militias.
The Boing Boing Store’s Gift Guide is full of ideas for pretty much anyone in your life like hipster ice cub trays, Xbox controllers, Halo Boards, and even diamond necklaces. As always, all products in the Boing Boing Store come at great discounts, too. Shop by price bucket starting at under $20. Under $20:Bloxx Jumbo Ice Trays […]
Unlike traditional lighters, the SaberLight features an electronic plasma beam that’s both rechargeable and butane-free. This sleek lighter is even approved by TSA, so you’ll never be stuck buying lighters you’ll just have to throw away partially used. For some people, like me, this is a pretty big game-changer. The SaberLight’s beam is actually both hotter and cleaner […]
Holiday shopping is in full swing, and the Striiv Touch is one of the best gift ideas I’ve landed on. Its simple design works for females and males, and its wide range of features makes it suitable for even the non-fitness enthusiasts in your life.Unlike traditional fitness trackers, the Striiv Touch also acts as a smartwatch. It […]