Relaxation: Evaluation and Applications

Fekete, Gyorgy; Eklundh, Jan‐Olof; Rosenfeld, Azriel

doi:10.1109/tpami.1981.4767131

Cited by 43 publications

(9 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…One of the first techniques was Rosenfeld's [63] who used local average. We also have the use of relaxation [64,65], the Laplacian of the images [24], quadtrees [66] and second-order statistics [67].…”

Section: Global Thresholding Methods and Neural Network In Image Segmentioning

confidence: 99%

Global Image Thresholding Adaptive Neuro-Fuzzy Inference System Trained with Fuzzy Inclusion and Entropy Measures

Bogiatzis

Papadopoulos

2019

Symmetry

View full text Add to dashboard Cite

Thresholding algorithms segment an image into two parts (foreground and background) by producing a binary version of our initial input. It is a complex procedure (due to the distinctive characteristics of each image) which often constitutes the initial step of other image processing or computer vision applications. Global techniques calculate a single threshold for the whole image while local techniques calculate a different threshold for each pixel based on specific attributes of its local area. In some of our previous work, we introduced some specific fuzzy inclusion and entropy measures which we efficiently managed to use on both global and local thresholding. The general method which we presented was an open and adaptable procedure, it was free of sensitivity or bias parameters and it involved image classification, mathematical functions, a fuzzy symmetrical triangular number and some criteria of choosing between two possible thresholds. Here, we continue this research and try to avoid all these by automatically connecting our measures with the wanted threshold using some Artificial Neural Network (ANN). Using an ANN in image segmentation is not uncommon especially in the domain of medical images. However, our proposition involves the use of an Adaptive Neuro-Fuzzy Inference System (ANFIS) which means that all we need is a proper database. It is a simple and immediate method which could provide researchers with an alternative approach to the thresholding problem considering that they probably have at their disposal some appropriate and specialized data.

show abstract

Section: Global Thresholding Methods and Neural Network In Image Segmentioning

confidence: 99%

Global Image Thresholding Adaptive Neuro-Fuzzy Inference System Trained with Fuzzy Inclusion and Entropy Measures

Bogiatzis

Papadopoulos

2019

Symmetry

View full text Add to dashboard Cite

show abstract

“…1, return the best isomorphism f between two graphs. For instance: probabilistic relaxation [23], Graduated-Assignment [24] or Expectation-Maximisation [25]. In fact, the input of these algorithms can be matrices and instead of graphs and since matrices capture all the differences between graphs and the minimisation cost is defined through these matrices (eq.…”

Section: Graph Matching and Isomorphism Between Graphsmentioning

confidence: 99%

“…Algorithm Interactive Graph Matching presented in [31] obtains a labelling between nodes of attributed graphs and considering the human feedback. That is, it computes several times a sub-optimal graph-matching algorithm (for instance [23,24,25]), but in each step, the cost matrices and are modified through the current user feedback. In fact, we assume the input of the algorithm is not both graphs but matrices and .…”

Section: Active Algorithmmentioning

confidence: 99%

Active Graph Matching Based on Pairwise Probabilities between Nodes

Cortés

Serratosa

Solé-Ribalta

2012

Lecture Notes in Computer Science

View full text Add to dashboard Cite

We propose a method to perform active graph matching in which the active learner queries one of the nodes of the first graph and the oracle feedback is the corresponding node of the other graph. The method uses any graph matching algorithm that iteratively updates a probability matrix between nodes (Graduated Assignment, Expectation Maximisation or Probabilistic Relaxation). The oracle's feedback is used to update the costs between nodes and arcs of both graphs. We present and validate four different active strategies based on the probability matrix between nodes. It is not needed to modify the code of the graph-matching algorithms, since our method simply needs to read the probability matrix and to update the costs between nodes and arcs. Practical validation shows that with few oracle's feedbacks, the algorithm finds the labelling that the user considers optimal because imposing few labellings the other ones are corrected automatically.

show abstract

“…First, for each pair of AGs, an assignation matrix is computed. To do so, several error-tolerant graph matching algorithms have been presented, such as, probabilistic relaxation [11], softassign [5] or Expectation-Maximisation [12]. Each cell of the assignation matrix M ai stores the probability of node a, from G 1 , to be assigned to node i, from graph G 2 , that is, the probability of the labelling ( ) The cost to obtain this matrix, using softassign [5], 4 ) per iteration of the algorithm.…”

Section: Consistency Indexmentioning

confidence: 99%

“…However, both procedures are generic enough to be applied to any other algorithms that use a probabilistic approach, e.g. [8] and [11].…”

Section: Is O((n/2-n)·rmentioning

confidence: 99%

On the Computation of the Common Labelling of a Set of Attributed Graphs

Solé-Ribalta

Serratosa

2009

Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications

View full text Add to dashboard Cite

Abstract. In some methodologies, it is needed a consistent common labelling between the vertices of a set of graphs, for instance, to compute a representative of a set of graphs. This is a NP-problem with an exponential computational cost depending on the number of nodes and the number of graphs. The aim of this paper is twofold. On one hand, we aim to establish a technical methodology to define this problem for the present and further research. On the other hand, we present two sub-optimal algorithms to compute the labelling between a set of graphs. Results show that our new algorithms are able to find a consistent common labelling while reducing, most of the times, the mean distance of the AG set.

show abstract

Relaxation: Evaluation and Applications

Cited by 43 publications

References 7 publications

Global Image Thresholding Adaptive Neuro-Fuzzy Inference System Trained with Fuzzy Inclusion and Entropy Measures

Global Image Thresholding Adaptive Neuro-Fuzzy Inference System Trained with Fuzzy Inclusion and Entropy Measures

Active Graph Matching Based on Pairwise Probabilities between Nodes

On the Computation of the Common Labelling of a Set of Attributed Graphs

Contact Info

Product

Resources

About