With the rapid development of modern video technologies, the concept of "Big Brother's all-seeing eye" ceases to be something abstract: every day millions of people become objects of shooting numerous covert surveillance cameras. Unfortunately, it will soon become even more difficult to hide from them.
The fact is that the researchers of the Boston University laboratory under the leadership of Vivek Goyal have developed a computer algorithm that allows any video cameras to literally "look around the corner."
The method was named "computational periscope". It is based on the analysis of shadows that objects cast on walls, for which powerful decoding algorithms have been developed. So far, the resulting image is not perfect, but it is already quite distinguishable.
Everything would be much easier if there were mirrored walls, but since their surface is matte, scattering light in all directions, the reflected image is a blurred spot.
To get the original image of the object located around the corner, the program needs an image of the wall on which the shadow is reflected from it. It is worth clarifying - the object should not be transparent. In fact, the algorithm uses complex calculations to decode light.
One of the main advantages of the technology is the ability to use it when defining an image captured with any digital camera. Now it takes 48 seconds for the program to recreate the image "around the corner" using a regular computer. Computers with higher power will take even less time.
In a practical sense, the new technology can be in demand in the creation of unmanned vehicles and autonomous robots operating in natural disaster zones.