Results 1  10
of
22
Learning compatibility coefficients for relaxation labeling processes
 IEEE Trans. Pattern Anal. Machine Intell
, 1994
"... AbstractRelaxation labeling processes have been widely used in many different domains including image processing, pattern recognition, and artificial intelligence. They are iterative procedures that aim at reducing local ambiguities and achieving global consistency through a parallel exploitation o ..."
Abstract

Cited by 41 (5 self)
 Add to MetaCart
AbstractRelaxation labeling processes have been widely used in many different domains including image processing, pattern recognition, and artificial intelligence. They are iterative procedures that aim at reducing local ambiguities and achieving global consistency through a parallel exploitation of contextual information, which is quantitatively expressed in terms of a set of “compatibility coefficients. ” The problem of determining compatibility coefficients has received a considerable attention in the past and many heuristic, statisticalbased methods have been suggested. In this paper, we propose a rather different viewpoint to solve this problem: we derive them attempting to optimize the performance of the relaxation algorithm over a sample of training data; no statistical interpretation is given: compatibility coefficients are simply interpreted as real numbers, for which performance is optimal. Experimental results over a novel application of relaxation are given, which prove the effectiveness of the proposed approach. Index Terms Compatibility coefficients, constraint satisfaction, gradient projection, learning, neural networks, nonlinear
Local Consistency in Parallel ConstraintSatisfaction Networks
 Artificial Intelligence
, 1994
"... We summarize our work on the parallel complexity of local consistency in constraint networks, and present several basic techniques for achieving parallel execution of constraint networks. We are interested primarily in developing a classification of constraint networks according to whether they admi ..."
Abstract

Cited by 11 (1 self)
 Add to MetaCart
We summarize our work on the parallel complexity of local consistency in constraint networks, and present several basic techniques for achieving parallel execution of constraint networks. We are interested primarily in developing a classification of constraint networks according to whether they admit massively parallel execution. The major result supported by our investigations is that the parallel complexity of constraint networks is critically dependent on subtle properties of the network that do not influence its sequential complexity. 1 Introduction In this position paper we summarize our work on the parallel complexity of local consistency in constraint networks [Kas90, Kas86, Kas89, KRS87, KD90]. Our research is aimed at deriving a precise characterization of the utility of parallelism in such networks. We are interested primarily in developing a classification of constraint networks according to whether they admit massively parallel execution. We have analyzed parallel executio...
A Relaxation Algorithm for Realtime Multiple View 3DTracking
 Image and Vision Computing
, 2002
"... this paper we present a discrete relaxation algorithm for reducing the intrinsic combinatorial complexity by pruning the decision tree based on unreliable prior information from independent 2Dtracking for each view. The algorithm improves the reliability of spatiotemporal correspondence by simulta ..."
Abstract

Cited by 7 (0 self)
 Add to MetaCart
this paper we present a discrete relaxation algorithm for reducing the intrinsic combinatorial complexity by pruning the decision tree based on unreliable prior information from independent 2Dtracking for each view. The algorithm improves the reliability of spatiotemporal correspondence by simultaneous optimisation over multiple views in the case where 2Dtracking in one or more views are ambiguous. Application to the 3D reconstruction of human movement, based on tracking of skincoloured regions in three views, demonstrates considerable improvement in reliability and performance. Results demonstrate that the optimisation over multiple views gives correct 3D reconstruction and object labeling in the presence of incorrect 2Dtracking whilst maintaining realtime performance
Discriminative learning of maxsum classifiers
 Journal of Machine Learning Research
"... The maxsum classifier predicts ntuple of labels from ntuple of observable variables by maximizing a sum of quality functions defined over neighbouring pairs of labels and observable variables. Predicting labels as MAP assignments of a Random Markov Field is a particular example of the maxsum cla ..."
Abstract

Cited by 7 (1 self)
 Add to MetaCart
(Show Context)
The maxsum classifier predicts ntuple of labels from ntuple of observable variables by maximizing a sum of quality functions defined over neighbouring pairs of labels and observable variables. Predicting labels as MAP assignments of a Random Markov Field is a particular example of the maxsum classifier. Learning parameters of the maxsum classifier is a challenging problem because even computing the response of such classifier is NPcomplete in general. Estimating parameters using the Maximum Likelihood approach is feasible only for a subclass of maxsum classifiers with an acyclic structure of neighbouring pairs. Recently, the discriminative methods represented by the perceptron and the Support Vector Machines, originally designed for binary linear classifiers, have been extended for learning some subclasses of the maxsum classifier. Besides the maxsum classifiers with the acyclic neighbouring structure, it has been shown that the discriminative learning is possible even with arbitrary neighbouring structure provided the quality functions fulfill some additional constraints. In this article, we extend the discriminative approach to other three classes of maxsum classifiers with an arbitrary neighbourhood structure. We derive learning algorithms for two subclasses of maxsum classifiers whose response can be computed in polynomial time: (i) the maxsum classifiers with supermodular quality functions and (ii) the maxsum classifiers whose response can be computed exactly by a linear programming relaxation. Moreover, we show that the learning problem can be approximately solved even for a general maxsum classifier.
An algorithm using lengthr paths to approximate subgraph isomorphism
 PATTERN
, 2003
"... The ‘LeRP’ algorithm approximates subgraph isomorphism for attributed graphs based on counts of LengthR Paths. The algorithm provides a good approximation to the maximal isomorphic subgraph. The basic approach of the LeRP algorithm differs fundamentally from other methods. When comparing structura ..."
Abstract

Cited by 5 (2 self)
 Add to MetaCart
The ‘LeRP’ algorithm approximates subgraph isomorphism for attributed graphs based on counts of LengthR Paths. The algorithm provides a good approximation to the maximal isomorphic subgraph. The basic approach of the LeRP algorithm differs fundamentally from other methods. When comparing structural similarity LeRP uses a neighborhood of nodes that varies in size dynamically. This approach provides sufficient evidence of similarity to permit LeRP to form a nodetonode mapping and can be computed with polynomial effort in the worstcase. Results from over 32,000 simulated cases are reported. We demonstrate that LeRP does not need a high dynamic range of node and edge coloring to perform well. For example, LeRP can input 50node and 100node graphs that contain a common 50node subgraph, and then compute a matching subgraph having 49.74 +/ 0.46 nodes (mean +/ one standard deviation). This takes from 0.4 to 0.5 seconds. In this example, 100 trials were evaluated and graphs had discrete coloring for nodes and edges with a dynamic range of four. Test conditions are varied and include strongly regular graphs as well as Model A.
The Rapidly Deployable Radio Network
 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS
, 1999
"... The Rapidly Deployable Radio Network (RDRN) is an architecture and experimental system to develop and evaluate hardware and software components suitable for implementing mobile, rapidly deployable, and adaptive wireless communications systems. The driving application for the RDRN is the need to quic ..."
Abstract

Cited by 4 (0 self)
 Add to MetaCart
The Rapidly Deployable Radio Network (RDRN) is an architecture and experimental system to develop and evaluate hardware and software components suitable for implementing mobile, rapidly deployable, and adaptive wireless communications systems. The driving application for the RDRN is the need to quickly establish a communications infrastructure following a natural disaster, during a law enforcement activity, or rapid deployment of military force. The RDRN project incorporates digitally controlled antenna beams, programmable radios, adaptive protocols at the link layer, and mobile node management. This paper describes the architecture for the Rapidly Deployable Radio Network and a prototype system built to evaluate key system components.
LeRP: An Algorithm Using LengthR Paths To Determine Subgraph Isomorphism
 Pattern Rec Journal
, 2001
"... The LeRP algorithm determines subgraph isomorphism for attributed graphs based on counts of LengthR Paths. The algorithm provides a good approximation to the maximal isomorphic subgraph. The paradigm associated with the LeRP algorithm differs fundamentally from other approaches. When comparing str ..."
Abstract

Cited by 3 (2 self)
 Add to MetaCart
(Show Context)
The LeRP algorithm determines subgraph isomorphism for attributed graphs based on counts of LengthR Paths. The algorithm provides a good approximation to the maximal isomorphic subgraph. The paradigm associated with the LeRP algorithm differs fundamentally from other approaches. When comparing structural similarity it uses a neighborhood of nodes, which varies in size dynamically. This approach provides sufficient evidence of similarity to permit LeRP to form a nodetonode mapping in just a few iterations. LeRP requires polynomial effort for each of these iterations. And, just three iterations were used for all of the 32,000 simulated trials reported herein. Results from an image registration application are also presented.
Quadratic Minimization for Labeling Problems
, 2002
"... Many tasks in Computer Vision can be formulated in the framework of Labeling Problems. Thereby we are asked to assign labels to objects. The assignment is based on a prior model for observationals in the sehensfeld and posteriori data. The labeling is to compute which minimizes ambiguities in the me ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
(Show Context)
Many tasks in Computer Vision can be formulated in the framework of Labeling Problems. Thereby we are asked to assign labels to objects. The assignment is based on a prior model for observationals in the sehensfeld and posteriori data. The labeling is to compute which minimizes ambiguities in the measurements. This computation involves an appropriate functional over objects and labels, which defines a notion of consistency.