The printed circuit board layout inspection methods are mostly based on local geometric information, therefore it is well suited to the cellular neural network (CNN) paradigm. Two layout errors are detected here namel...
详细信息
The printed circuit board layout inspection methods are mostly based on local geometric information, therefore it is well suited to the cellular neural network (CNN) paradigm. Two layout errors are detected here namely, the breaks in the wires and some kind of short circuits. The designed analogic algorithms to solve the problems above were tested on real life examples using an experimental system based on our CNN-HAC1M digital multiprocessor add-on-board, with 1 million cell space and 2.0 /spl mu/s/cell/iteration speed.
The cellular neural network (CNN) Universal Machine architecture when implemented in VLSI chips needs special interfaces to provide for efficient system performance both in time (speed) and space (image size). The CNN...
详细信息
The cellular neural network (CNN) Universal Machine architecture when implemented in VLSI chips needs special interfaces to provide for efficient system performance both in time (speed) and space (image size). The CNN chip set architectures described solve this problem, including interfacing the possibly analog input sensors and digital output and control. Various forms of CNN Engines are presented embedding CNN chip sets. A new device, the Visual Mouse, a hand held visual supercomputer, is also presented which exploits the genuine features of CNN Engines.
The architecture of ACE, a multiprocessor analogic cellular neural network (CNN) emulator engine consisting of 2 to 16 TMS320C40 floating point DSPs is introduced. The engine containing up to 512 Mbyte RAM (enough to ...
详细信息
The architecture of ACE, a multiprocessor analogic cellular neural network (CNN) emulator engine consisting of 2 to 16 TMS320C40 floating point DSPs is introduced. The engine containing up to 512 Mbyte RAM (enough to store a 512/spl times/512/spl times/512 sized CNN cube) which can be controlled through its SCSI port. It can either accelerate the multilayer CNN simulator CNNM or be accessed directly from the high level, C-based analogic CNN language ACL to achieve the simulation speed of /spl sim/2.8 /spl mu/sec/cell/iteration/DSP for 3/spl times/3 linear templates.
Size modification of binary pictures can be mapped onto the CNN array using space variant linear templates. However, if all the parameters have to be set for each cell individually, then one of the CNN's main adva...
详细信息
Size modification of binary pictures can be mapped onto the CNN array using space variant linear templates. However, if all the parameters have to be set for each cell individually, then one of the CNN's main advantages will be lost in practice, the simple and quick parallel reprogrammability. In this paper, a general methodology is presented to derive the space variant templates of the complete weighting matrix from control pictures applying a simple nonlinear space invariant template. The straightforward design method presumes a modified CNN architecture (multiple input and specific nonlinear voltage-controlled current sources in every cell) and can be extended for a large class of sparse weighting matrices. Following this strategy the diminishment and enlargement process has been investigated using constant cell current and various bias maps in the transformations.
Recognition of small patterns covering only a few pixels in an image cannot be done by conventional recognition methods. A theoretically new pattern recognition method has been developed for undersampled objects which...
详细信息
Recognition of small patterns covering only a few pixels in an image cannot be done by conventional recognition methods. A theoretically new pattern recognition method has been developed for undersampled objects which are (much) smaller than the window-size of a picture element (pixel), i.e. these objects are of subpixel size. The proposed statistical technique compares the gray-level histogram of the patterns of a set of scanned objects to be examined with the (calculated) gray-level densities of different (in shape or size) possible objects, and the recognition is based on this comparison. This method does not need high-precision movement of scanning sensors or any additional hardware. Moreover, the examined patterns should be randomly distributed on the screen, or a random movement of camera is (or target or both are) needed. Effects of noise are analysed, and filtering processes are suggested in the histogram domain. Several examples of different object shapes (triangle, rectangle, square, circle, curving lines, etc.) are presented through simulations and experiments. A number of possible application areas are suggested, including astronomy, line-drawing analysis and industrial laser measurements.
A multilayer analogic CNN (cellular neural network) algorithm is proposed to reconstruct surfaces from real scene stereo-pairs consisting of two images shot from one and the same camera position. The whole scene is il...
详细信息
A multilayer analogic CNN (cellular neural network) algorithm is proposed to reconstruct surfaces from real scene stereo-pairs consisting of two images shot from one and the same camera position. The whole scene is illuminated by randomly structured light to provide texture on otherwise "empty" surfaces. One image is that of the illuminated background, the other one is the image of illuminated objects in front of the same background.< >
This paper introduces an analytic method to determine the sensitivity to random parameter variations of analog VLSI neural network architectures for linear image filtering. The authors compare the robustness of severa...
详细信息
This paper introduces an analytic method to determine the sensitivity to random parameter variations of analog VLSI neural network architectures for linear image filtering. The authors compare the robustness of several different circuit architectures for low pass filtering. This method can also determine which components within a particular architecture should specified the most precisely.< >
In this paper it is shown how the Cellular neural Network (CNN) can be used to perform image and volume deblurring, with particular emphases on applications to microscopy. We discuss the basic linear theory of the CNN...
详细信息
In this paper it is shown how the Cellular neural Network (CNN) can be used to perform image and volume deblurring, with particular emphases on applications to microscopy. We discuss the basic linear theory of the CNN including issues of stability and template size. It is observed that a CNN with a small template can be used to implement an Infinite Impulse Response filter. It is then shown how general deblurring problems can be addressed with a CNN when the blurring operator is known. The proposed application is to solve the basic 3-D confocal image reconstruction task about the form of the blurring operator, confocal behavior in microscope images can be obtained with only 3-5 acquired image planes. In addition, the stored program capability of the CNN Universal Machine would provide integration of several image processing and detection tasks in the same architecture.< >
In this paper we report on a fast, complex and efficient implementation of the Cellular neural Network Universal Machine as an IC chip. The chip has continuous time analog dynamics, and has been designed to process 50...
详细信息
In this paper we report on a fast, complex and efficient implementation of the Cellular neural Network Universal Machine as an IC chip. The chip has continuous time analog dynamics, and has been designed to process 500,000 image frames per second.< >
Novel types of analogic algorithms, using spatio-temporal CNN (cellular nonlinear/neural networks) operations are introduced. These algorithms make complex decisions in images without reading out the CNN chip. This ma...
详细信息
Novel types of analogic algorithms, using spatio-temporal CNN (cellular nonlinear/neural networks) operations are introduced. These algorithms make complex decisions in images without reading out the CNN chip. This makes them extremely time, area, and power effective. Two crucial effects are emphasized: diffusion type templates are applied during a finite time interval and local logic operates within well defined parts (patches) in the image plane. Hence, a new type of pattern recognition algorithm is introduced. The technique is demonstrated on an example. In our example we are dealing with an actual problem: how to avoid the counterfeiting on color copiers.< >
暂无评论