G06V 10/762
Definition
Diese Klassifikationsstelle umfasst:(Für diese Definition ist die deutsche Übersetzung noch nicht abgeschlossen)
Techniques of grouping patterns together in order to reveal a certain structure or a meaning in images or video.
Notes – technical background
These notes provide more information about the technical subject matter that is classified in this place:
The object of techniques classified here is to identify groups of similar entities and to assign entities to a group (cluster) according to a measure of their similarity. Separability is determined by measuring the similarity or dissimilarity. Such techniques are usually performed in a high-dimensional feature space constructed by extracting features from the image or video, but can also be performed in the original domain, e.g. in the image domain in the case of image segmentation.
Regarding the grouping of patterns, any pattern may belong exclusively to a single cluster (hard clustering) or it may belong simultaneously to more than one cluster up to a certain degree (fuzzy clustering) according to a similarity (or proximity) measure. In addition, depending on the clustering method used, proximity may be defined (a) between vectors, (b) between a vector and a set of vectors (or a cluster), and (c) between sets of vectors (or different clusters).
Examples of proximity measures are: dissimilarity measures (based on l1, l2, and l∞ norms), similarity measures (inner product, cosine, Pearson’s correlation coefficient, Tanimoto distance, etc.).
Clustering algorithms include:
a) clustering based on statistical measures (which mainly employ numerical data) which adopt a cost function J related to possible groupings which is subject to a global or local optimisation criterion, and return a clustering that optimises J. Examples of such algorithms are:
- Hard clustering algorithms, where a vector belongs exclusively to a specific cluster, e.g. k-means, k-medoids, Linde-Buzo-Gray, ISODATA, DBSCAN, Neural Gas;
- Fuzzy clustering algorithms, where a vector belongs to a specific cluster up to a certain degree, e.g. fuzzy c-means, adaptive fuzzy C-shells (AFCS), fuzzy C quadric shells (FCQS), modified fuzzy C quadric shells (MFCQS);
- Probabilistic clustering algorithms, which follow Bayesian classification arguments and in which each vector is assigned to the cluster according to a probabilistic set-up, e.g. expectation maximisation (EM), Gaussian mixture model (GMM), mean-shift;
b) Graph-based clustering algorithms, e.g. minimum spanning tree (MST) clustering, clustering based on directed trees, spectral clustering, graph-cut optimisation;
c) Competitive learning algorithms for clustering, in which a set of representatives is selected and the goal is to move all representatives to regions of a vector space that are “dense” in terms of other vectors. Examples are leaky learning algorithms, self-organising maps (SOM), learning vector quantisation (LVQ).
Hierarchical clustering is a popular technique in the class of graph-based clustering, with its agglomerative or divisive variants. Various criteria can be used for determining the groupings, such as those based on matrix theory involving dissimilarity matrices. Algorithms included in this scheme are:
- Single link algorithm;
- Complete link algorithm;
- Weighted pair group method average (WPGMA);
- Unweighted pair group method average (UPGMA);
- Weighted pair group method centroid (WPGMC);
- Ward or minimum variance algorithm.
Examples
Clustering face images to detect affinity between persons using a graph-based clustering algorithm
Querverweise
Informative Querverweise
| Pattern recognition or machine learning, using classification | G06V 10/764 |
| Pattern recognition or machine learning, using regression | G06V 10/766 |
| Pattern recognition or machine learning, processing image features in feature spaces | G06V 10/77 |
| Pattern recognition or machine learning, fusion | G06V 10/80 |
| Information retrieval of still images; Clustering; Classification | G06F 16/55 |
| Information retrieval of video data; Clustering; Classification | G06F 16/75 |
| Image analysis; Segmentation; Edge detection | G06T 7/10 |
Glossar
G06V 10/762
Definition Statement
This place covers:Techniques of grouping patterns together in order to reveal a certain structure or a meaning in images or video.
Notes – technical background
These notes provide more information about the technical subject matter that is classified in this place:
The object of techniques classified here is to identify groups of similar entities and to assign entities to a group (cluster) according to a measure of their similarity. Separability is determined by measuring the similarity or dissimilarity. Such techniques are usually performed in a high-dimensional feature space constructed by extracting features from the image or video, but can also be performed in the original domain, e.g. in the image domain in the case of image segmentation.
Regarding the grouping of patterns, any pattern may belong exclusively to a single cluster (hard clustering) or it may belong simultaneously to more than one cluster up to a certain degree (fuzzy clustering) according to a similarity (or proximity) measure. In addition, depending on the clustering method used, proximity may be defined (a) between vectors, (b) between a vector and a set of vectors (or a cluster), and (c) between sets of vectors (or different clusters).
Examples of proximity measures are: dissimilarity measures (based on l1, l2, and l∞ norms), similarity measures (inner product, cosine, Pearson’s correlation coefficient, Tanimoto distance, etc.).
Clustering algorithms include:
a) clustering based on statistical measures (which mainly employ numerical data) which adopt a cost function J related to possible groupings which is subject to a global or local optimisation criterion, and return a clustering that optimises J. Examples of such algorithms are:
- Hard clustering algorithms, where a vector belongs exclusively to a specific cluster, e.g. k-means, k-medoids, Linde-Buzo-Gray, ISODATA, DBSCAN, Neural Gas;
- Fuzzy clustering algorithms, where a vector belongs to a specific cluster up to a certain degree, e.g. fuzzy c-means, adaptive fuzzy C-shells (AFCS), fuzzy C quadric shells (FCQS), modified fuzzy C quadric shells (MFCQS);
- Probabilistic clustering algorithms, which follow Bayesian classification arguments and in which each vector is assigned to the cluster according to a probabilistic set-up, e.g. expectation maximisation (EM), Gaussian mixture model (GMM), mean-shift;
b) Graph-based clustering algorithms, e.g. minimum spanning tree (MST) clustering, clustering based on directed trees, spectral clustering, graph-cut optimisation;
c) Competitive learning algorithms for clustering, in which a set of representatives is selected and the goal is to move all representatives to regions of a vector space that are “dense” in terms of other vectors. Examples are leaky learning algorithms, self-organising maps (SOM), learning vector quantisation (LVQ).
Hierarchical clustering is a popular technique in the class of graph-based clustering, with its agglomerative or divisive variants. Various criteria can be used for determining the groupings, such as those based on matrix theory involving dissimilarity matrices. Algorithms included in this scheme are:
- Single link algorithm;
- Complete link algorithm;
- Weighted pair group method average (WPGMA);
- Unweighted pair group method average (UPGMA);
- Weighted pair group method centroid (WPGMC);
- Ward or minimum variance algorithm.
Examples
Clustering face images to detect affinity between persons using a graph-based clustering algorithm
References
Informative references
| Pattern recognition or machine learning, using classification | G06V 10/764 |
| Pattern recognition or machine learning, using regression | G06V 10/766 |
| Pattern recognition or machine learning, processing image features in feature spaces | G06V 10/77 |
| Pattern recognition or machine learning, fusion | G06V 10/80 |
| Information retrieval of still images; Clustering; Classification | G06F 16/55 |
| Information retrieval of video data; Clustering; Classification | G06F 16/75 |
| Image analysis; Segmentation; Edge detection | G06T 7/10 |