IEEE Transactions on Pattern Analysis and Machine Intelligence, 1988
This paper deals with the problem of digital visual inspection of thin-film disk heads. We will p... more This paper deals with the problem of digital visual inspection of thin-film disk heads. We will present machine vision algorithms and a supporting architecture that are integrated in a fully automated prototype system for disk head inspection. We will also elaborate on some specific methods, such as computation of the Hough transform and multicode masks in pipeline architectures, object segmentation in textured backgrounds, and matching of extracted defects with inspection specifications. Extensive experimented results will also be given.
Journal of The Optical Society of America A-optics Image Science and Vision, 1986
In this paper, a review of machine-vision methods for microelectronic production and related unso... more In this paper, a review of machine-vision methods for microelectronic production and related unsolved problems is given, based on research paradigms and manufacturing needs that are seldom addressed in the literature. Although the importance of many industrial machine-vision applications has been identified, this paper will cover only a subset of these problems because of space limitations. Specifically, the automated visual inspection of printed circuit boards and thick-film circuits will be reviewed. Different aspects of these problems will be surveyed, including manufacturing considerations, contributions to sensing techniques, and, more importantly, digital image processing and analysis methods. For printed wiring boards, several production stages are addressed in which automated visual inspection plays a key role, from artwork to lamination of multiple layers and plated-through holes. In addition, the main inspection problems arising in populated boards will be reviewed, such as solder joint integrity and top-surface component placement. Finally, some machine-vision approaches used for thick-film and hybrid circuit inspection are considered.
Journal of Parallel and Distributed Computing, 1987
This paper deals with a novel architecture that makes real-time projection-based image processing... more This paper deals with a novel architecture that makes real-time projection-based image processing a reality. The design is founded on raster-mode processing, which is exploited in a powerful and flexible pipeline. This architecture, dubbed ''P³E'' (Parallel Pipeline Projection Engine), supports a large variety of image processing and image analysis applications. In the present paper, the authors concern themselves with several
IEEE Transactions on Pattern Analysis and Machine Intelligence, 1987
In this correspondence, some image transforms and features such as projections along linear patte... more In this correspondence, some image transforms and features such as projections along linear patterns, convex hull approximations, Hough transform for line detection, diameter, moments, and principal components will be considered. Specifically, we present algorithms for computing these features which are suitable for implementation in image analysis pipeline architectures. In particular, random access memories and other dedicated hardware components which may be found in the implementation of classical techniques are not longer needed in our algorithms. The effectiveness of our approach is demonstrated by running some of the new algorithms in conventional short-pipelines for image analysis. In related papers, we have shown a pipeline architecture organization called PPPE (Parallel Pipeline Projection Engine), which unleashes the power of projection-based computer vision, image processing, and computer graphics. In the present correspondence, we deal with just a few of the many algorithms which can be supported in PPPE. These algorithms illustrate the use of the Radon transform as a tool for image analysis.
Two parallel algorithms are presented for the problem of labeling the connected components of a b... more Two parallel algorithms are presented for the problem of labeling the connected components of a binary image. The machine model is an SIMD two-dimensional mesh-connected computer consisting of an N×N array of processing elements, each containing a single pixel of an N×N image. Both new algorithms use a local shrinking operation defined by S. Levialdi (1972) and have time complexities of O(N log N) bit operations, making them the fastest local algorithms for the problem. Compared to other approaches with similar or better asymptotic time complexities, this local approach greatly simplifies the algorithms and reduces the constants of proportionality by nearly two orders of magnitude, making them the first practical algorithms for the problem. The two algorithms differ in the amount of memory required per processing element; the first uses O(N) bits, while the second uses a novel compression scheme to reduce the requirement to O(log N ) bits
IEEE Transactions on Pattern Analysis and Machine Intelligence, 1989
Abstmct-In this paper, an important mid-level task for computer vision is addressed. The problem ... more Abstmct-In this paper, an important mid-level task for computer vision is addressed. The problem consists of labeling connected components in X NI/* binary images. This task can be solved with parallel computers by using a simple and novel algorithm.
This paper presents O(log' N) time algorithms for labeling the connected components of an N'i2 x ... more This paper presents O(log' N) time algorithms for labeling the connected components of an N'i2 x N'12 pixel binary image using an N processor hypercube or shuffle-exchange computer. The algorithms that are presented are the first to solve this problem in O(log' N) time using the given models of parallel computers. The algorithms are based on a divide-and-conquer approach and use as a subroutine an O(log N) time PRAM algorithm for labeling the connected components of a graph. The simulation of the PRAM by the hypercube and shuffle-exchange computers is particularly efficient because the PRAM that is being simulated has only cI(N~'~) processors and memory cells. 0 1989 AC.&IICC PXSS, IX
This paper presents a parallel sorting algorithm called Cubesort. Cubesort sorts N data items by ... more This paper presents a parallel sorting algorithm called Cubesort. Cubesort sorts N data items by performing a number of rounds, each of which partitions the N data items into groups of size S and sorts within the groups. For many values of N and S, Cubesort requires fewer such rounds than are required by any previously published algorithm. Cubesort can also be used to sort N data items on hypercube, shuffle-exchange, and cube-connected cycles computers with P processors in time U(N log* N/P log(N/P)) over a wide range of the parameters N and P. In particular, when N = Plrl/' and k is a constant, Cubesort sorts on the above parallel computers in O(N log N/P) time, thus obtaining an optimal processortime product for comparison sorting. The application of Cubesort to general routing problems is also discussed. o
The problem of Fourier-transform phase reconstruction from the Fourier-transform magnitude of mul... more The problem of Fourier-transform phase reconstruction from the Fourier-transform magnitude of multidimensional discrete signals is considered. It is well known that, if a discrete finite-extent n-dimensional signal (n > 2) has an irreducible z transform, then the signal is uniquely determined from the magnitude of its Fourier transform. It is also known that this irreducibility condition holds for all multidimensional signals except for a set of signals that has measure zero. We show that this uniqueness condition is stable in the sense that it is not sensitive to noise. Specifically, it is proved that the set of signals whose z transform is reducible is contained in the zero set of a certain multidimensional polynomial. Several important conclusions can be drawn from this characterization, and, in particular, the zero-measure property is obtained as a simple byproduct.
We deal with iterative least-squares solutions of the linear signal-restoration problem g = Af. F... more We deal with iterative least-squares solutions of the linear signal-restoration problem g = Af. First, several existing techniques for solving this problem with different underlying models are unified. Specifically, the following are shown to be special cases of a general iterative procedure [H. Bialy, Arch. Ration. Mech. Anal. 4, 166 (1959)] for solving linear operator equations in Hilbert spaces: (1) a Van Cittert-type algorithm for deconvolution of discrete and continuous signals; (2) an iterative procedure for regularization when g is contaminated with noise; a Papoulis-Gerchberg algorithm for extrapolation of continuous signals [A. Papoulis, IEEE Trans. Circuits Syst. CAS-22, 735 (1975); R. W. Gerchberg, Opt. Acta 21, 709 (1974)]; (4) an iterative algorithm for discrete extrapolation of band-limited infinite-extent discrete signals land the minimum-norm property of the extrapolation obtained by the iteration [A. Jain and S. Ranganath, IEEE Trans. Acoust. Speech Signal Process. ASSP-29, (1981)fl; and (5) a certain iterative procedure for extrapolation of band-limited periodic discrete signals [V. Tom et al., IEEE Trans.
IEEE Transactions on Pattern Analysis and Machine Intelligence, 1989
... often spurred by dif-ferent motivations, has approached the problem from a variety of angles,... more ... often spurred by dif-ferent motivations, has approached the problem from a variety of angles, and has contributed to a number of the-oretical and experimental results 181-12 I]. Although important progress was reported in [8], a gen-eral extension of Logan's results to 2 and ...
IEEE Transactions on Pattern Analysis and Machine Intelligence, 1988
This paper deals with the problem of digital visual inspection of thin-film disk heads. We will p... more This paper deals with the problem of digital visual inspection of thin-film disk heads. We will present machine vision algorithms and a supporting architecture that are integrated in a fully automated prototype system for disk head inspection. We will also elaborate on some specific methods, such as computation of the Hough transform and multicode masks in pipeline architectures, object segmentation in textured backgrounds, and matching of extracted defects with inspection specifications. Extensive experimented results will also be given.
Journal of The Optical Society of America A-optics Image Science and Vision, 1986
In this paper, a review of machine-vision methods for microelectronic production and related unso... more In this paper, a review of machine-vision methods for microelectronic production and related unsolved problems is given, based on research paradigms and manufacturing needs that are seldom addressed in the literature. Although the importance of many industrial machine-vision applications has been identified, this paper will cover only a subset of these problems because of space limitations. Specifically, the automated visual inspection of printed circuit boards and thick-film circuits will be reviewed. Different aspects of these problems will be surveyed, including manufacturing considerations, contributions to sensing techniques, and, more importantly, digital image processing and analysis methods. For printed wiring boards, several production stages are addressed in which automated visual inspection plays a key role, from artwork to lamination of multiple layers and plated-through holes. In addition, the main inspection problems arising in populated boards will be reviewed, such as solder joint integrity and top-surface component placement. Finally, some machine-vision approaches used for thick-film and hybrid circuit inspection are considered.
Journal of Parallel and Distributed Computing, 1987
This paper deals with a novel architecture that makes real-time projection-based image processing... more This paper deals with a novel architecture that makes real-time projection-based image processing a reality. The design is founded on raster-mode processing, which is exploited in a powerful and flexible pipeline. This architecture, dubbed ''P³E'' (Parallel Pipeline Projection Engine), supports a large variety of image processing and image analysis applications. In the present paper, the authors concern themselves with several
IEEE Transactions on Pattern Analysis and Machine Intelligence, 1987
In this correspondence, some image transforms and features such as projections along linear patte... more In this correspondence, some image transforms and features such as projections along linear patterns, convex hull approximations, Hough transform for line detection, diameter, moments, and principal components will be considered. Specifically, we present algorithms for computing these features which are suitable for implementation in image analysis pipeline architectures. In particular, random access memories and other dedicated hardware components which may be found in the implementation of classical techniques are not longer needed in our algorithms. The effectiveness of our approach is demonstrated by running some of the new algorithms in conventional short-pipelines for image analysis. In related papers, we have shown a pipeline architecture organization called PPPE (Parallel Pipeline Projection Engine), which unleashes the power of projection-based computer vision, image processing, and computer graphics. In the present correspondence, we deal with just a few of the many algorithms which can be supported in PPPE. These algorithms illustrate the use of the Radon transform as a tool for image analysis.
Two parallel algorithms are presented for the problem of labeling the connected components of a b... more Two parallel algorithms are presented for the problem of labeling the connected components of a binary image. The machine model is an SIMD two-dimensional mesh-connected computer consisting of an N×N array of processing elements, each containing a single pixel of an N×N image. Both new algorithms use a local shrinking operation defined by S. Levialdi (1972) and have time complexities of O(N log N) bit operations, making them the fastest local algorithms for the problem. Compared to other approaches with similar or better asymptotic time complexities, this local approach greatly simplifies the algorithms and reduces the constants of proportionality by nearly two orders of magnitude, making them the first practical algorithms for the problem. The two algorithms differ in the amount of memory required per processing element; the first uses O(N) bits, while the second uses a novel compression scheme to reduce the requirement to O(log N ) bits
IEEE Transactions on Pattern Analysis and Machine Intelligence, 1989
Abstmct-In this paper, an important mid-level task for computer vision is addressed. The problem ... more Abstmct-In this paper, an important mid-level task for computer vision is addressed. The problem consists of labeling connected components in X NI/* binary images. This task can be solved with parallel computers by using a simple and novel algorithm.
This paper presents O(log' N) time algorithms for labeling the connected components of an N'i2 x ... more This paper presents O(log' N) time algorithms for labeling the connected components of an N'i2 x N'12 pixel binary image using an N processor hypercube or shuffle-exchange computer. The algorithms that are presented are the first to solve this problem in O(log' N) time using the given models of parallel computers. The algorithms are based on a divide-and-conquer approach and use as a subroutine an O(log N) time PRAM algorithm for labeling the connected components of a graph. The simulation of the PRAM by the hypercube and shuffle-exchange computers is particularly efficient because the PRAM that is being simulated has only cI(N~'~) processors and memory cells. 0 1989 AC.&IICC PXSS, IX
This paper presents a parallel sorting algorithm called Cubesort. Cubesort sorts N data items by ... more This paper presents a parallel sorting algorithm called Cubesort. Cubesort sorts N data items by performing a number of rounds, each of which partitions the N data items into groups of size S and sorts within the groups. For many values of N and S, Cubesort requires fewer such rounds than are required by any previously published algorithm. Cubesort can also be used to sort N data items on hypercube, shuffle-exchange, and cube-connected cycles computers with P processors in time U(N log* N/P log(N/P)) over a wide range of the parameters N and P. In particular, when N = Plrl/' and k is a constant, Cubesort sorts on the above parallel computers in O(N log N/P) time, thus obtaining an optimal processortime product for comparison sorting. The application of Cubesort to general routing problems is also discussed. o
The problem of Fourier-transform phase reconstruction from the Fourier-transform magnitude of mul... more The problem of Fourier-transform phase reconstruction from the Fourier-transform magnitude of multidimensional discrete signals is considered. It is well known that, if a discrete finite-extent n-dimensional signal (n > 2) has an irreducible z transform, then the signal is uniquely determined from the magnitude of its Fourier transform. It is also known that this irreducibility condition holds for all multidimensional signals except for a set of signals that has measure zero. We show that this uniqueness condition is stable in the sense that it is not sensitive to noise. Specifically, it is proved that the set of signals whose z transform is reducible is contained in the zero set of a certain multidimensional polynomial. Several important conclusions can be drawn from this characterization, and, in particular, the zero-measure property is obtained as a simple byproduct.
We deal with iterative least-squares solutions of the linear signal-restoration problem g = Af. F... more We deal with iterative least-squares solutions of the linear signal-restoration problem g = Af. First, several existing techniques for solving this problem with different underlying models are unified. Specifically, the following are shown to be special cases of a general iterative procedure [H. Bialy, Arch. Ration. Mech. Anal. 4, 166 (1959)] for solving linear operator equations in Hilbert spaces: (1) a Van Cittert-type algorithm for deconvolution of discrete and continuous signals; (2) an iterative procedure for regularization when g is contaminated with noise; a Papoulis-Gerchberg algorithm for extrapolation of continuous signals [A. Papoulis, IEEE Trans. Circuits Syst. CAS-22, 735 (1975); R. W. Gerchberg, Opt. Acta 21, 709 (1974)]; (4) an iterative algorithm for discrete extrapolation of band-limited infinite-extent discrete signals land the minimum-norm property of the extrapolation obtained by the iteration [A. Jain and S. Ranganath, IEEE Trans. Acoust. Speech Signal Process. ASSP-29, (1981)fl; and (5) a certain iterative procedure for extrapolation of band-limited periodic discrete signals [V. Tom et al., IEEE Trans.
IEEE Transactions on Pattern Analysis and Machine Intelligence, 1989
... often spurred by dif-ferent motivations, has approached the problem from a variety of angles,... more ... often spurred by dif-ferent motivations, has approached the problem from a variety of angles, and has contributed to a number of the-oretical and experimental results 181-12 I]. Although important progress was reported in [8], a gen-eral extension of Logan's results to 2 and ...
Uploads
Papers by Jorge Sanz