GPU Startup Story: Camero Goes Beyond Seeing Through Walls to Real-Time Body ScanningOctober 30, 2012
Mild-mannered, understated and neatly groomed, Ron Daisy could well be Israel’s answer to Clark Kent.
But, unlike Superman’s alter ego, the signal-processing PhD doesn’t need a phone booth to access the superpower to see through walls. He helped invent it, as Camero-Tech’s research chief.
His Tel Aviv area startup – which presented at NVIDIA’s recent Israel GPU symposium – pioneered technology that uses short-range radar based on ultra-wide-band microwave technology to enable SWAT teams, police and the military in more than 20 countries identify concealed threats.
This technology is now the basis for the firm’s latest major project, real-time 3D security body scanning.
Still in the development phase, it would enable airports, for example, to scan passengers on the move, searching for hidden weapons. This updates the most recently introduced practice that requires them, one-by-one, to enter a transparent capsule, raise their hands and freeze for a few seconds as machinery moves around them.
“In time, this could be rolled out in thousands of sites,” said Daisy, at the head office of the 30-person homeland-security firm, which was acquired earlier this year by SK Group, an Israeli defense and security company. “It can scan multiple individuals simultaneously, without lag or moving parts. And it protects the privacy of those being scanned because it singles out concealed objects without displaying the person’s physical image.”
As a result, passengers – or individuals attending a concert, political event or other gathering – could simply proceed down a corridor and be automatically scanned. The solution, which leverages highly sophisticated algorithms processed on CUDA-enabled GPUs, would be faster, less costly, less invasive and far more portable than prevailing techniques.
Daisy proudly describes Camero’s new technology in a large room also used to show its “through-wall” technology. Parked along the periphery of the space are a wide range of props, including a half-dozen long slices of heavy walls mounted on wheels for easy maneuvering.
There’s an eight-foot section of tightly compressed mud and straw bricks commonly used in rural fighting areas. Another wall-on-wheels is made of large gray cinder blocks common in the Middle East. Others are constructed of combinations of plaster, wooden studs, wallboard and siding, replicating structures used in the U.S and Western Europe.
“We can see through any wall that isn’t continuous metal and detect what’s moving on the other side,” he said, pointing to a screen that shows images in 3D to a depth of 20 meters beyond the other side of the wall.
But today, Daisy and his colleague Hagay Keller, Camero’s VP of business development, whose background is in the defense industry, are keen to show their new scanning product.
Still in an early form, it comprises three long, broad rectangular panels, set at angles reminiscent of mirrors in a sci-fi tailor’s shop (see image, left). Mounted on each are four smaller boards containing irregular patterns of what appears to be mounted copper-colored butterflies. This set of 192 transmitters and a similar number of receivers is used to scatter and collect microwaves emitted in micro-bursts that consume only a few milliwatts, less energy than a cell phone.
The received data is transmitted using two one-gigabyte lines to the system’s back end, where the data gets crunched by a server fitted out with four NVIDIA-based GPUs. These reconstruct and process the scanned image, and then render it in 3D, showing details smaller than one centimeter. While weapons and other objects are immediately visible, the system can’t distinguish between, say, a smartphone and an explosive device, so it would need to be used in conjunction with other security techniques.
Daisy and Keller expect the device to be commercially deployed in the near future. They note that it has significant applications beyond homeland security. The low-cost, real-time scanning capabilities have applications in the medical field, such as cancer detection, as well as industrial settings, where it could be deployed to search for product defects.