Menu

Timesdelhi.com

May 23, 2019
Category archive

Computer Vision

Reality Check: The marvel of computer vision technology in today’s camera-based AR systems

in animation/AR/ar/vr/Artificial Intelligence/Augmented Reality/Column/Computer Vision/computing/Delhi/Developer/digital media/Gaming/gif/Global Positioning System/gps/India/mobile phones/neural network/Politics/starbucks/TC/Virtual Reality/VR by

British science fiction writer, Sir Arther C. Clark, once said, “Any sufficiently advanced technology is indistinguishable from magic.”

Augmented reality has the potential to instill awe and wonder in us just as magic would. For the very first time in the history of computing, we now have the ability to blur the line between the physical world and the virtual world. AR promises to bring forth the dawn of a new creative economy, where digital media can be brought to life and given the ability to interact with the real world.

AR experiences can seem magical but what exactly is happening behind the curtain? To answer this, we must look at the three basic foundations of a camera-based AR system like our smartphone.

  1. How do computers know where it is in the world? (Localization + Mapping)
  2. How do computers understand what the world looks like? (Geometry)
  3. How do computers understand the world as we do? (Semantics)

Part 1: How do computers know where it is in the world? (Localization)

Mars Rover Curiosity taking a selfie on Mars. Source: https://www.nasa.gov/jpl/msl/pia19808/looking-up-at-mars-rover-curiosity-in-buckskin-selfie/

When NASA scientists put the rover onto Mars, they needed a way for the robot to navigate itself on a different planet without the use of a global positioning system (GPS). They came up with a technique called Visual Inertial Odometry (VIO) to track the rover’s movement over time without GPS. This is the same technique that our smartphones use to track their spatial position and orientation.

A VIO system is made out of two parts.

News Source = techcrunch.com

Edgybees’s new developer platform brings situational awareness to live video feeds

in argus/Augment/Augmented Reality/Computer Vision/Delhi/Developer/Edgybees/Emerging-Technologies/Enterprise/gps/India/jerusalem/Marketing/OurCrowd/Politics/san diego/Startups/Technology by

San Diego-based Edgybees today announced the launch of Argus, its API-based developer platform that makes it easy to add augmented reality features to live video feeds.

The service has long used this capability to run its own drone platform for first responders and enterprise customers, which allows its users to tag and track objects and people in emergency situations, for example, to create better situational awareness for first responders.

I first saw a demo of the service a year ago, when the team walked a group of journalists through a simulated emergency, with live drone footage and an overlay of a street map and the location of ambulances and other emergency personnel. It’s clear how these features could be used in other situations as well, given that few companies have the expertise to combine the video footage, GPS data and other information, including geographic information systems, for their own custom projects.

Indeed, that’s what inspired the team to open up its platform. As the Edgybees team told me during an interview at the Ourcrowd Summit last month, it’s impossible for the company to build a new solution for every vertical that could make use of it. So instead of even trying (though it’ll keep refining its existing products), it’s now opening up its platform.

“The potential for augmented reality beyond the entertainment sector is endless, especially as video becomes an essential medium for organizations relying on drone footage or CCTV,” said Adam Kaplan, CEO and co-founder of Edgybees. “As forward-thinking industries look to make sense of all the data at their fingertips, we’re giving developers a way to tailor our offering and set them up for success.”

In the run-up to today’s launch, the company already worked with organizations like the PGA to use its software to enhance the live coverage of its golf tournaments.

News Source = techcrunch.com

Blind users can now explore photos by touch with Microsoft’s Seeing AI

in accessibility/Apps/Artificial Intelligence/Augmented Reality/Blindness/Computer Vision/Delhi/Disabilities/India/machine learning/Microsoft/mobile/Politics by

Microsoft’s Seeing AI is an app that lets blind and limited-vision folks convert visual data into audio feedback, and it just got a useful new feature. Users can now use touch to explore the objects and people in photos.

It’s powered by machine learning, of course, specifically object and scene recognition. All you need to do is take a photo or open one up in the viewer and tap anywhere on it.

“This new feature enables users to tap their finger to an image on a touch-screen to hear a description of objects within an image and the spatial relationship between them,” wrote Seeing AI lead Saqib Shaikh in a blog post. “The app can even describe the physical appearance of people and predict their mood.”

Because there’s facial recognition built in as well, you could very well take a picture of your friends and hear who’s doing what and where, and whether there’s a dog in the picture (important) and so on. This was possible on an image-wide scale already, as you can see in this image:

But the app now lets users tap around to find where objects are — obviously important to understanding the picture or recognizing it from before. Other details that may not have made it into the overall description may also appear on closer inspection, such as flowers in the foreground or a movie poster in the background.

In addition to this, the app now natively supports the iPad, which is certainly going to be nice for the many people who use Apple’s tablets as their primary interface for media and interactions. Lastly, there are a few improvements to the interface so users can order things in the app to their preference.

Seeing AI is free — you can download it for iOS devices here.

News Source = techcrunch.com

Brodmann17 nabs $11M for its automotive computer vision tech that runs on any CPU

in Artificial Intelligence/Automotive/autonomous cars/brodmann17/Computer Vision/Delhi/India/Politics/self-driving cars by

Efficient computer vision systems are a critical component of autonomous and assisted driving vehicles, and now a startup that has developed a way to deliver computer vision technology without relying on costly and bulky hardware — by building deep learning software that can run even on low-end processors — has secured a round of funding as it gears up for its first services later this year.

Brodmann17 — named after the primary visual cortex in the human brain — has raised $11 million in a Series A round of funding led by OurCrowd, with participation also from Maniv Mobility, AI Alliance, UL Ventures, Samsung NEXT, and the Sony Innovation Fund.

Brodmann17’s edge technology (necessary for rapid computations) is designed to be used in any automotive application that requires artificial intelligence to see and process objects, roads and wider landscapes, competing with the likes of Intel’s Mobileye, services being developed by other OEMs like Bosch, and some automakers like BMW.

The challenge that all these and many other players in the self-driving industry are tackling is that as cars start to be viewed more as hardware, they are taking on some of the biggest feature challenges ever tackled in the world of tech. Autonomous systems are not only expensive but they consume a lot of energy and take up a lot of space in the vehicle, so everyone is looking for solutions that can be less of a strain in one or ideally all of these areas.

The pitch from Brodmann17 is that its core product essentially does just that: its deep-learning based computer vision technology is designed as a “light weight” solution, which can work even on smaller, low-end processors. (Note: it works on low-end CPUs but not nearly as well as on the faster ones.)

The plan is for Brodmann17’s tech to eventually be a part of fully autonomous deployments, but with self-driving vehicles still some years away from being a reality, CEO Adi Pinhas — a deep learning and computer vision specialist who co-founded the company with two other AI scientists, Amir Alush and Assaf Mushinsky — said that its first commercial efforts will come in the form of advanced driver assistance systems (ADAS): it’s currently working with a global, tier-one automaker to incorporate its technology into front and rear cameras to make more accurate identifications of still and moving objects when a human is still behind the wheel.

That is not a small fish: ADAS is not only already a key component of many newer vehicles, but its ubiquity and functionality will continue to grow. ADAS systems — which are often supplied in part or entirety by third parties to carmakers — was a $20 billion market in 2017 and is projected to reach nearly $92 billion by 2025.

I first met the founding team of Broadmann in Tel Aviv, where they are based, a couple of years ago, when it was just four guys working in a corner of the Samsung NEXT incubator in the city, showing me the the earliest versions of how its tech was able to sit on small processors and identify with a lot of nuance different small and large objects that are encountered in a typical driving scenario.

Fast forward to today, and the company now has 70 people, mostly engineers, now working out of its own digs, and the startup is continuing to hire as it gears up beyond that early development phase.

Pinhas said that in these last couple of years, he’s seen some interesting evolutions in how the tech world, and the wider automotive industry, have approached the concept of autonomous cars.

On the one hand, everyone is throwing what they can at self-driving, and that will inevitably help accelerate roadmaps for new prototypes and tests. On the other hand, that increased effort is also leading to more pragmatism on just how much work lies ahead and how more elements of self driving might come sooner than full-fledged systems.

“Right now, to me it looks like the market may have taken a step back. Everyone wants to speed up development on autonomous systems, but at the same time I noticed how this year at CES, no one was talking about Level 5,” Pinhas said, referring to the highest level of autonomy in driving services, and the big tech event in January where many of the next big shiny new services get shown off for the first time. “I think the thinking now is, even a working Level 4 deployment would be great. Let’s do that and see how well we can get robo-taxis driving in well-defined scenarios.”

That’s where Brodmann17’s push into ADAS comes in: it gives the company a foothold in future deployments and services while also proving the concept by powering services that are live today. 

The other interesting development that Pinhas noted is a shift in focus from volumes of training data, to developing smarter neural networks to calculate and understand that data. “It used to be ‘who has more data’, but now everyone has it,” he said. “Now it’s about the algorithms for training. Experts have long thought that neural networks” — designed to “think” like humans — “will solve everything but the key is still figuring out how best to train those networks. Just throwing data at them will not solve this.” Notably, this is an area where Broadmann17 has put focus for some time, “and others are also starting to now.”

Pinhas admits that Mobileye is the most advanced company in the automotive computer vision market today, although we are still at such an early stage of development that there is room for more innovation, and more startups and other large companies to make an impact. This is why investors are interested in Brodmann17, and why the startup has already started working on its next round to supply itself with capital for the next phase.

“We are convinced that Brodmann17 is one of the best deep AI companies out there. The company has a very experienced management team and exceptional technical leadership that has created a major leapfrog in the fundamentals of AI algorithms,” said Eli Nir, Senior Partner at OurCrowd, said. “Brodmann17’s technology opens the doors for low computation implementation of AI – significantly lowering cost, complexity, and price, and can be used over many sectors and industries. We are very excited to lead this round and take part in the future success of the company.”

News Source = techcrunch.com

Koala-sensing drone helps keep tabs on drop bear numbers

in Artificial Intelligence/Australia/Computer Vision/conservation/Delhi/drones/Gadgets/Hardware/India/machine learning/Politics/Science/TC/UAVs by

It’s obviously important to Australians to make sure their koala population is closely tracked — but how can you do so when the suckers live in forests and climb trees all the time? With drones and AI, of course.

A new project from Queensland University of Technology combines some well-known techniques in a new way to help keep an eye on wild populations of the famous and soft marsupials. They used a drone equipped with a heat-sensing camera, then ran the footage through a deep learning model trained to look for koala-like heat signatures.

It’s similar in some ways to an earlier project from QUT in which dugongs — endangered sea cows — were counted along the shore via aerial imagery and machine learning. But this is considerably harder.

A koala

“A seal on a beach is a very different thing to a koala in a tree,” said study co-author Grant Hamilton in a news release, perhaps choosing not to use dugongs as an example because comparatively few know what one is.

“The complexity is part of the science here, which is really exciting,” he continued. “This is not just somebody counting animals with a drone, we’ve managed to do it in a very complex environment.”

The team sent their drone out in the early morning, when they expected to see the greatest contrast between the temperature of the air (cool) and tree-bound koalas (warm and furry). It traveled as if it was a lawnmower trimming the tops of the trees, collecting data from a large area.

Infrared image, left, and output of the neural network highlighting areas of interest

This footage was then put through a deep learning system trained to recognize the size and intensity of the heat put out by a koala, while ignoring other objects and animals like cars and kangaroos.

For these initial tests, the accuracy of the system was checked by comparing the inferred koala locations with ground truth measurements provided by GPS units on some animals and radio tags on others. Turns out the system found about 86 percent of the koalas in a given area, considerably better than an “expert koala spotter,” who rates about a 70. Not only that, but it’s a whole lot quicker.

“We cover in a couple of hours what it would take a human all day to do,” Hamilton said. But it won’t replace human spotters or ground teams. “There are places that people can’t go and there are places that drones can’t go. There are advantages and downsides to each one of these techniques, and we need to figure out the best way to put them all together. Koalas are facing extinction in large areas, and so are many other species, and there is no silver bullet.”

Having tested the system in one area of Queensland, the team is now going to head out and try it in other areas of the coast. Other classifiers are planned to be added as well, so other endangered or invasive species can be identified with similar ease.

Their paper was published today in the journal Nature Scientific Reports.

News Source = techcrunch.com

1 2 3 5
Go to Top