Showing papers by "José María Carazo published in 1994"
TL;DR: In this work, a classification of both aligned but heterogeneous image data sets as well as basically homogeneous but otherwise rotationally misaligned image populations is performed, in the latter case completely avoiding the typical reference dependency of correlation-based alignment methods.
108 citations
TL;DR: A modification of guided self-scheduling as well as efficient buffering strategies are presented in a new parallel computing approach to the case of exhaustive sequential sequence comparison algorithms on message-passing architectures.
Abstract: We present a new parallel computing approach to the case of exhaustive sequential sequence comparison algorithms on message-passing architectures. In this context a modification of guided self-scheduling as well as efficient buffering strategies are presented. We discuss two specific implementations, one on the Paramid parallel computer, and the other on a cluster of workstations running PVM. In both cases the parallel performance is higher than with any other method presented so far. The code is public domain and can be obtained by anonymous ftp at ftp.cnb.uam.es.
20 citations
18 Apr 1994
TL;DR: It will be shown how communication latency and overall data-passing load is significantly reduced, as well as dynamic load balancing and fault tolerant capabilities are achieved.
Abstract: In this work we discuss possible strategies for mapping sequential sequence comparison algorithms onto LAN-based message passing computer architectures The software-integration tool used is the public-domain PVM 31 system We present two possible mapping strategies, one centered around a file server using Guided Self Scheduling and the other based on socket to socket communications We will show how communication latency and overall data-passing load is significantly reduced, as well as dynamic load balancing and fault tolerant capabilities are achieved The mapping methodology used in this study can scale well from simple workstation clusters to far more complex general message-passing architectures
9 citations
TL;DR: Experimental results on the use of the Bispectrum for translational invariant image averaging and a number of practical studies on how many Bispectra from noisy realizations of a given image are needed for a meaningful image recovery as a function of the image signal-to-noise ratio are presented.
4 citations
02 Jan 1994
1 citations
09 Sep 1994
TL;DR: A working prototype of a distributed data base of volume data from different forms of 3D imaging that is sufficiently general to be applicable to data in a number of different fields of biomedical science.
Abstract: In this communication we present a working prototype of a distributed data base of volume data from different forms of 3D imaging. Examples of 3D information both from electron tomography of biological macromolecules and from medical imaging are already included in this development. This type of structural information is crosslinked, whenever applicable, to other sources of information, such as bibliographies. The solution we present here is sufficiently general to be applicable to data in a number of different fields of biomedical science.
1 citations
26 Jan 1994
TL;DR: This work treats the particular problem of the three dimensional reconstruction of a 3D object from 2D images and develops a parallel algorithm that is general in the sense that it does not impose any restrictions on the size of the problem and is notdependent on the dimensions of the mesh.
Abstract: The filtered backprojection algorithm is a popularmethod for the reconstruction of n-dimensionalsignals from their (n-1) dimensional projections. Inthis work we will treat the particular problem of thethree dimensional reconstruction of a 3D object from2D images. The parallel algorithm we develop isgeneral in the sense that it does not impose anyrestrictions on the size of the problem and is notdependent on the dimensions of the mesh. We havealso compared this algorithm in multiprocessorarchitectures with hypercube and mesharrangements, obtaining similar computation times.This last aspect indicates that the hypercubeincreases the hardware cost with no significantimprovement in the efficiency of the algorithms. 1: Sequential algorithm. The foundation on which all three dimensionalreconstruction techniques are based was originallydeveloped by Radon in 1917 [1]. He analyzed what is nowknown as the Randon Transform and its inverse.Currently, three dimensional macromoleculereconstruction algorithms can be grouped into two largesets: On one hand we find the series expansion methods,among which ART (Algebraic Reconstruction Technique)and its many variants stand out. On the other we have________________________________________