Slideshare uses cookies to improve functionality and performance, and to provide you with relevant advertising. If you continue browsing the site, you agree to the use of cookies on this website. See our User Agreement and Privacy Policy. See our Privacy Policy and User Agreement for details. Published on Apr 1, SlideShare Explore Search You.

Author:Gonris Faegis
Language:English (Spanish)
Published (Last):8 July 2011
PDF File Size:4.17 Mb
ePub File Size:1.92 Mb
Price:Free* [*Free Regsitration Required]

To browse Academia. Skip to main content. By using our site, you agree to our collection of information through the use of cookies. To learn more, view our Privacy Policy. Log In Sign Up. Manu Ravi. Combine brain-power with seeing power, and you have the fastest, cheapest, most extra ordinary processor ever-the human eye.

Little wonder, research labs the world over are striving to produce a near-perfect electronic eye. The GVPP, which crunches 20 billion instructions per second BIPS , models the human perceptual process at the hardware level by mimicking the separate temporal and spatial functions of the eye-to-brain system.

The processor sees its environment as a stream of histograms regarding the location and velocity of objects. GVPP has been demonstrated as capable of learning-in-place to solve a variety of pattern recognition problems. It boasts automatic normalization for varying object size, orientation and lighting conditions, and can function in daylight or darkness.

This electronic "eye" on a chip can now handle most tasks that a normal human eye can. That includes driving safely, selecting ripe fruits, reading and recognizing things. Sadly, though modeled on the visual perception capabilities of the human brain, the chip is not really a medical marvel, poised to cure the blind.

Such devices can be termed an electronic spatio-temporal neuron, and are particularly useful for image processing, but may also be used for other signals, such as audio signals. The techniques of the present invention are also particularly useful for tracking one or more objects in real time. It is desirable to provide devices including combined data processing units of a similar nature, each addressing a particular parameter extracted from the video signal.

In particular, it is desirable to provide devices including multiple units for calculating histograms, or electronic spatio-temporal neuron STN, each processing a DATA A , by a function in order to generate individually an output value.

The present invention also provides a method for perception of an object using characteristics, such as its shape, its size or its orientation, using a device composed of a set of histogram calculation units. Using the techniques of the present invention, a general outline of a moving object is determined with respect to a relatively stable background, then inside this outline, elements that are characterized by their tone, color, relative position etc.

It would be relatively simple for a CMOS chip to implement in hardware the separate contributions of temporal and spatial processing in the brain. The brain-eye system uses layers of parallel-processing neurons that pass the signal through a series of preprocessing steps, resulting in real-time tracking of multiple moving objects within a visual scene.

Pirim created a chip architecture that mimicked the work of the neurons, with the help of multiplexing and memory. That means it can track an object through varying light sources or changes in size, as when an object gets closer to the viewer or moves farther away. The GVPP on the other hand, adapt to real time changes in lighting without recalibration, day or light. For many decades the field of computing has been trapped by the limitations of the traditional processors. Many futuristic technologies have been bound by limitations of these processors.

These limitations stemmed from the basic architecture of these processors. Traditional processors work by slicing each and every complex program into simple tasks that a processor could execute.

This requires an existence of an algorithm for solution of 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 performs well. It can solve a problem with its neural learning function. Neural networks are extremely fault tolerant.

By their design even if a group of neurons get, the neural network only suffers a smooth degradation of the performance. This is a crucial difference, from traditional processors as they fail to work even if a few components are damaged.

GVPP recognizes stores , matches and process patterns. Even if pattern is not recognizable to a human programmer in input the neural network, it will dig it out from the input. Thus GVPP becomes an efficient tool for applications like the pattern matching and recognition. The basic element here is a neuron. There are large number of input lines and an output line to a neuron. Each neuron is capable of implementing a simple function.

It takes the weighted sum of its inputs and produces an output that is fed into the next layer. The weights assigned to each input are a variable quantity. A large number of such neurons interconnected form a neural network.

Every input that is given to the neural network gets transmitted over entire network via direct connections called synaptic connections and feed back paths. Thus the signal ripples in the neural network, every time changing the weighted values associated with each input of every neuron.

These changes in the ripples will naturally direct the weights to modify into those values that will become stable. That is, those values does not change. At this point the information about the signal is stored as the weighted values of inputs in the neural network. A neural network geometrizes computation. When we draw the state diagram of a neural network, the network activity burrows a trajectory in this state space.

The trajectory begins with a computation problem. The problem specifies initial conditions which define the beginning of trajectory in the state space.

In pattern learning, the pattern to be learned defines the initial conditions. Where as in pattern recognition, the pattern to be recognized defines the initial conditions. Most of the trajectory consists of transient behavior or computations. The weights associated with inputs gradually change to learn new pattern information. The trajectory ends when the system reaches equilibrium. This is the final state of the neural network.

If the pattern was meant to be matched, the final neuronal state represents the pattern that is closest match to the input pattern. Video signal S is composed of a succession of frames, wherein each frame includes a succession of pixels whose assembly forms a space, for example an image for a two-dimensional space.

Real-time outputs perceive, recognize and analyze both static images and time- varying patterns for specific objects, their heading, speed, shading and color differences. By mimicking the eye and the visual regions of the brain, the GVPP puts together the salient features necessary for recognition.

So instead of capturing frames of pixels, the chip identifies objects of interest, determines each object's speed and direction, then follows them by tracking their color through the scene. The chip emulates the eye, which has 5 million cones sensitive to color, only 15 per cent of which see "blue" the rest are red and green and million monochromatic rods that are 35 times more sensitive than the cones through two processing steps-tonic and phasic.

Tonic processing auto-scales according to ambient light conditions, enabling it to adapt to a range of luminosity. Phasic processing determines movement by using local variable in the feedback loops loops. The eye typically sweeps a scene about two to three times a second as well as making vibratory movements at about Hz.

The faster jitter accounts for the eye-sensitivity to the smallest detectable feature, which is an edge moving between adjacent rods or cones. After all this processing the visual signal is then sent to the brain for higher-level observation and recognition tasks.

Because only detected movements and color along with the shape and contour of objects is sent to the brain, rather than raw pixels, the average compression ratio information is about A sensor array is a set of several sensors that an information gathering device uses to gather information usually directional in nature that cannot be gathered from a single source for a central processing unit. Each pixel is read by the vision chip with hardware that determines and scales luminescence, tracks color, remembers movement in the previous moment, recalls the direction of previous movement, and then deduces the speed of the various detected objects from parallel phasic and tonic neural circuitry.

Basically, each parameter has an associated neuron that handles its processing tasks in parallel. In addition, each pixel has two auxiliary neurons that define the zone in which an object is located-that is, from the direction in which an object is moving, these neurons deduce the leading and trailing edge of the object and mark with registers associated with the first leading-edge and last trailing-edge pixel belonging to the object.

Each of these silicon neurons is built with RAM, a few registers, an adder and a comparator. Supplied as a pin module, the chip accommodates analogue-input line levels for video input, with an input amplifier with programmable gain auto scaling the signal. The modules measure 40 square mm, have pins and can handle MHz video signals. BEV dubs the neural-learning capability of its development environment "programming by seeing and doing," because of its ease of use.

The engineer needs no knowledge of the internal workings of the GVPP, the company said, only application- specific domain knowledge. The system's modular approach permits the developer to create a hierarchy of application building blocks that simplify problems with inheritable software characteristics.

Based on the neural network learning functions, the chip allows application software to be developed quickly through a combination of simple operational commands and immediate feedback.

Simple applications such as detecting and tracking a single moving object can be programmed in less than a day. Even complex applications such as detecting a driver falling sleep can be programmed in a month. The ability of a neural network to learn from the data is perhaps its greatest ability.

This learning of neural network can be described as its configuration such that the application of a set of inputs produces the desired set of outputs. Various methods to set the strengths of the connections exist. One way is to set the synaptic connection explicitly, using a priori knowledge. The GVPP further subcategorizes pixels by ranges, for instance luminance within 10 percent and 65 percent, hue of blue, saturation between 20 and 25 percent, and moving upward in scene.

A set of second-level pattern recognition commands permits the GVPP to search for different objects in different parts of the scene — for instance, to look for a closed eyelid only within the rectangle bordered by the corners of the eye. Since some applications may also require multiple levels of recognition, the GVPP has software hooks to pass along the recognition task from level to level. For instance, to detect when a driver is falling asleep — a capability that could find use in California, which is about to mandate that cars sound an "alarm" when drowsy drivers begin to nod off — the GVPP is first programmed to detect the driver's head, for which it creates histograms of head movement.

The microprocessor reads these histograms to identify the area for the eye. Then the recognition task passes to the next level, which searches only within the eye area rectangles. High-speed movement there, normally indicative of blinking, is discounted, but when blinks become slower than a predetermined level, they are interpreted as the driver nodding off, and trigger an alarm. The GVPP multiplexes this neural hardware with off-chip scratchpad memory to simulate as many as , synaptic connections per neuron.

Each of these synapses can be changed through the on-chip microprocessor for a combined processing total of over 6.



Toggle navigation. Help Preferences Sign up Log in. View by Category Toggle navigation. Products Sold on our sister site CrystalGraphics. Tags: generic perception processor visual generic perception processor visual. Latest Highest Rated. The GVPP tracks an object anticipating where its leading and trailing edges makes differences with the background When an object gets closer to the viewer or moves farther away That it can track an object through varying light sources or changes in size 8 Major performance strength Adaptation to varying light sources -means GVPP adapt to real time changes in lighting without recalibration,day or light Limitation of traditional processors were removed -traditional processors slice each and every complex program into simple tasks -requires an algorithm GVPP does not require an algorithm Solve a problem using neural learning function Fault tolerent 9 HOW IT WORKS?



We think you have liked this presentation. If you wish to download it, please recommend it to your friends in any social system. Share buttons are a little bit lower. Thank you! Published by Gyles Owens Modified over 4 years ago.


Generic Visual Perception Processor (GVPP)

To browse Academia. Skip to main content. By using our site, you agree to our collection of information through the use of cookies. To learn more, view our Privacy Policy. Log In Sign Up. Manu Ravi. Combine brain-power with seeing power, and you have the fastest, cheapest, most extra ordinary processor ever-the human eye.


Skip to search form Skip to main content You are currently offline. Some features of the site may not work correctly. TheGeneric visual perception processor is a single chip modeled on the perception capabilities of the human brain, which can detect objects in a motion video signal and then locate and track them in real time. Imitating the human eyes neural networks and the brain, the chip can handle about 20 billion instructions per second. Save to Library. Create Alert.

Related Articles