Approximate, common sense reasoning (driving, playing piano, baseball player)
What is Neural Networks?
It is an interconnection of artificial neurons.
A set of nodes (units, neurons, processing elements)
Each node has input and output.
Each node performs a simple computation by its
Connectivity gives the structure/architecture of
What can be computed by a NN is primarily
determined by the connections and their
Why Cellular Neural Networks?
The major disadvantage of neural networks
implementations is the number of interconnections
In order to reduce the number of interconnections Chua &
Yang proposed Cellular Neural Networks where neurons
were only connected to other neurons within a certain
What is Cellular Neural Networks?
A Cellular Neural Network (CNN), also known as Cellular
Nonlinear Network, is an array of dynamical systems
or coupled networks with local connections only.
The network is an ensemble of spatially arranged cells,
where each cell is itself a dynamical system that is locally
coupled to its neighboring cells within some prescribed
History of CNN
In 1988 papers from Leon Chua & Lin Yang introduced the
concept of the Cellular Neural Network.
In 1993 Tamas Roska & Leon Chua’s article has introduced the
first analog CNN processor for the engineering research
One of the best review about CNN & its types and definitions
was given by Valerio Cimagalli & Marco Balsi.
Architecture of CNN
This is a two-dimensional cellular neural network.
The squares are the circuit units called cells.
The links between the cells indicate that there are interactions
between the linked cells.
The basic circuit unit of CNN is called a cell.
It contains linear & non-linear elements, which typicallyare linear capacitors, linear resistors, controlled sources & independent sources
Consider an M X N cellular neural network, having M X N cells arranged in M rows and N columns.
We call the cell on the i 'th row and the j’th column,
cell (i, j), and denote it by C(i, j) as in the above figure.
Now let us define what we mean by a neighborhood of C(i, j).
The r-neighborhood of a cell C(i, j) , in a cellular neural network is shown as
This figure shows 3 neighborhoods of the same cell with r=1,2 and 3 respectively.
It lends itself to local, low-level & processor intensive operations.
Image In painting
Moving Object Detection
Axis of Symmetry Detection
CNN processors can achieve upto 50,000 frames per second & for certain applications such as missile tracking, flash detection & sparkling plug diagnostics these processors have outperformed a conventional super computer.
CNN processors have implemented a real-time system that replicates mammalian retinas, validating that the original CNN modeled correct aspects of biological neural networks used to perform.
The Reconstructions Of Damaged Images Using Cellular Neural Network Model
The experimental results are obtained by using the "CadetWin" (CNN Application Development Environment and Toolkit under Windows)
Figure: Experimental results in different noise conditions with CNN.
Versatile in Nature
Easier to Integrate
CONCLUSION CNN processors have been implemented using current technology and there are plans to implement CNN processors into future technologies. They include the necessary interfaces for programming and interfacing, and have been implemented in a variety of systems and provide value to their users. Researchers are also transitioning CNN processors into emerging technologies.
Future Scope One potential future application of CNN microprocessors is to combine them with the DNA microarrays to allow for a near- real time DNA analysis of hundreds of thousands of different DNA sequences.
L. O. Chua and L. Yang, "Cellular neural networks: Applications,“ IEEE Trans. Circuits Syst., pp. 1273-1290, this issue.
J. J. Hopfield, "Neural networks and physical systems with emergent computational abilities," Proc. Natl. Acad. Sei. USA., vol. 79,pp. 2554-2558, 1982.
J. J. Hopfield and D. W. Tank "'Neural' computation of decisions in optimization problems," Biol. Cybern., vol. 52, pp. 141-152,1985.
I. Szatmari, P. Foldesy, C. Rekeczky and A. Zarandy, “Image Processing Library for the Aladdin Computer”, Int’l Workshop on Cellular Neural Networks and Their Applications
D. Balya, G, Tímar, G. Cserey, and T. Roska , “A New Computational Model for CNN-UMs and its Computational Complexity”, Int’l Workshop on Cellular Neural Networks and Their Applications