Trajectory of Dynamic Clusters in Image Time-Series

Héas, Patrick; Irit,; Dlr,; Datcu, Mihai; Giros, Alain

doi:10.1142/9789812702630_0005

Cited by 2 publications

(4 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On the first levels of the hierarchical modeling, strong families of models are applied to extract information using inference based on Bayesian and entropic methods. This unsupervised modeling results in a graph representation coding the information content of SITS [2]. The inferred graph ¢ characterizes cluster trajectories in the dynamic feature space related to multitemporal (MT) objects.…”

Section: Introductionmentioning

confidence: 99%

Bayesian Learning on Graphs for Reasoning on Image Time-Series

Héas¹

2004

AIP Conference Proceedings

View full text Add to dashboard Cite

Satellite image time-series (SITS) are multidimensional signals of high complexity. Their main characteristics are spatio-temporal patterns which describes the scene dynamics. The information contained in SITS was coded using Bayesian methods, resulting in a graph representation [2]. This paper further presents a concept of interactive learning for semantic labeling of spatio-temporal patterns present in SITS. It enables the recognition and the probabilistic retrieval of similar events. Graphs are attached to statistical models for spatio-temporal processes, which at their turn describe physical changes in the observed scene. Therefore, user-specific semantics attached to spatiotemporal events are modeled using combinations of parameters of a distance model between subgraphs. Thus, the learning step is performed by the incremental definition of a spatio-temporal event type via user-provided positive and negative sub-graph examples. From these examples we infer probabilities of the Bayesian network, based on a Dirichlet model, that links user interest to a specific similarity measurement. According to the current state of learning, sub-graph posterior probabilities are estimated. Experiments, performed on a multitemporal SPOT image time-series, demonstrate the presented reasoning concept.

show abstract

Section: Introductionmentioning

confidence: 99%

Bayesian Learning on Graphs for Reasoning on Image Time-Series

Héas¹

2004

AIP Conference Proceedings

View full text Add to dashboard Cite

show abstract

“…Therefore, the dynamic of image structures and spatial objects is analyzed in the multidimensional space. Figure 1 illustrates the analysis [2]. The signal is modeled and clustered in different FS representations.…”

Section: A Modeling and Clustering Multitemporal And Time-localized mentioning

confidence: 99%

“…Each Gaussian will be considered as a grouping of similar data points and thus, will define a cluster. In order to perform clustering without any constraints neither on the number of Gaussians present in the mixture nor on their parameters, a minimum description length criterion has been used to select the best model among all the possible Gaussian mixture configurations [2].…”

Section: A Modeling and Clustering Multitemporal And Time-localized mentioning

confidence: 99%

“…By associating the TL and the MT information, the problem of quantifying the distance between projected MT clusters can be reformulated in defining an information measurement, able to compare two different models and learn about their similarity. Using Kullback-Leibler divergence, which is relevant for these model comparisons, the dynamic FS modeling is inferred resulting in weighted MT cluster trajectories formalized trough TL evolution of clusters [2]. Therefore, this analysis results in a graph…”

Section: B Inference Of the Dynamic Feature Space Modelingmentioning

confidence: 99%

See 1 more Smart Citation

Image time-series mining

Héas

Marthon

Datcu

et al.

IEEE International IEEE International IEEE International Geoscience and Remote Sensing Symposium, 2004. IGARSS '04. Proceedings

Self Cite

View full text Add to dashboard Cite

Abstract-A visual information mining concept is proposed for spatio-temporal patterns discovery in remotely sensed image time-series (ITS). An information theory framework is adopted to first model information content. It results in the inference of a relevant directed graph characterizing ITS. Then the user conjecture is modeled via visual information representations : similarity measures between sub-graphs, which represents spatiotemporal events are derived and included in an interactive learning and probabilistic retrieval procedure of user-specific event-types. The present concept for ITS mining is demonstrated on multitemporal SPOT data.

show abstract

Trajectory of Dynamic Clusters in Image Time-Series

Abstract: Index Terms-Spatio-temporal learning, information mining, Bayesian modeling, dynamic cluster trajectories, semantic labeling.

Cited by 2 publications

References 7 publications

Bayesian Learning on Graphs for Reasoning on Image Time-Series

Bayesian Learning on Graphs for Reasoning on Image Time-Series

Image time-series mining

Contact Info

Product

Resources

About