05-10-2017, 03:53 PM
The 'generic visual perception processor (GVPP)' has been developed after 10 long years of scientific effort. The generic Visual Perception Processor (GVPP) can automatically detect objects and track their movement in real time.
GVPP, which crosses 20 billion instructions per second (BIPS), models the human perceptual process at the hardware level by imitating the temporal and spatial functions separate from the eye-brain system. The processor sees its surroundings as a stream of histograms with respect to the location and speed of objects.
GVPP has proven to be able to learn in-place to solve a variety of pattern recognition problems. It has automatic normalization to vary object size, orientation and lighting conditions, and can operate in daylight or darkness.
This electronic "eye" on a chip can now handle most of the tasks that a normal human eye can. That includes driving safely, selecting ripe fruits, reading and recognizing things. Sadly, although modeled on the visual perception capabilities of the human brain, the chip is not really a medical wonder, prepared to cure the blind.
GVPP tracks an "object," defined as a certain set of hue, luminance, and saturation values in a specific form, from frame to frame in a video sequence, anticipating where the address is and the output edges make "differences" with the background. This means that you can track an object through various light sources or changes in size, such as when an object approaches the viewer or moves further.
The greatest performance strength of GVPP over today's vision systems is its adaptation to varying light conditions. Today's vision systems dictate even shade less lighting, and even the next generation of prototype systems, designed to work under "normal" lighting conditions, can be used only at dawn at dusk. The GVPP, on the other hand, adapts to the changes in real time of the illumination without recalibration, day or light.
For many decades the field of computing has been trapped by the limitations of traditional processors. Many futuristic technologies have been limited by the limitations of these processors. These limitations come from the basic architecture of these processors. Traditional processors work by cutting out each complex program into simple tasks that a processor could execute. This requires the existence of an algorithm for solving the particular problem. But there are many situations where there is an inexistence of an algorithm or inability of a human to understand the algorithm.
Even in these extreme cases GVPP works fine. You can solve a problem with your neural learning function. Neural networks are extremely fault tolerant. By its design, even if a group of neurons obtain, the neural network only undergoes a mild degradation of the performance. You will not stop working abruptly. This is a crucial difference, from traditional processors, since they do not work, even if some of the components are damaged. GVPP recognizes stores, matches, and process patterns. Even if the pattern is not recognizable by a human programmer at the entrance of the neural network, it will dig it out of the input.
GVPP, which crosses 20 billion instructions per second (BIPS), models the human perceptual process at the hardware level by imitating the temporal and spatial functions separate from the eye-brain system. The processor sees its surroundings as a stream of histograms with respect to the location and speed of objects.
GVPP has proven to be able to learn in-place to solve a variety of pattern recognition problems. It has automatic normalization to vary object size, orientation and lighting conditions, and can operate in daylight or darkness.
This electronic "eye" on a chip can now handle most of the tasks that a normal human eye can. That includes driving safely, selecting ripe fruits, reading and recognizing things. Sadly, although modeled on the visual perception capabilities of the human brain, the chip is not really a medical wonder, prepared to cure the blind.
GVPP tracks an "object," defined as a certain set of hue, luminance, and saturation values in a specific form, from frame to frame in a video sequence, anticipating where the address is and the output edges make "differences" with the background. This means that you can track an object through various light sources or changes in size, such as when an object approaches the viewer or moves further.
The greatest performance strength of GVPP over today's vision systems is its adaptation to varying light conditions. Today's vision systems dictate even shade less lighting, and even the next generation of prototype systems, designed to work under "normal" lighting conditions, can be used only at dawn at dusk. The GVPP, on the other hand, adapts to the changes in real time of the illumination without recalibration, day or light.
For many decades the field of computing has been trapped by the limitations of traditional processors. Many futuristic technologies have been limited by the limitations of these processors. These limitations come from the basic architecture of these processors. Traditional processors work by cutting out each complex program into simple tasks that a processor could execute. This requires the existence of an algorithm for solving the particular problem. But there are many situations where there is an inexistence of an algorithm or inability of a human to understand the algorithm.
Even in these extreme cases GVPP works fine. You can solve a problem with your neural learning function. Neural networks are extremely fault tolerant. By its design, even if a group of neurons obtain, the neural network only undergoes a mild degradation of the performance. You will not stop working abruptly. This is a crucial difference, from traditional processors, since they do not work, even if some of the components are damaged. GVPP recognizes stores, matches, and process patterns. Even if the pattern is not recognizable by a human programmer at the entrance of the neural network, it will dig it out of the input.