Publications sur Image segmentation

Résultat de la recherche dans la liste des publications :

• 2 Articles
• 1 Article de conférence

nouvelle recherche | bibtex

affichage : 10 par page | 20 par page | 50 par page

2 Articles

1 - A Marked Point Process for Modeling Lidar Waveforms. C. Mallet et F. Lafarge et M. Roux et U. Soergel et F. Bretar et C. Heipke. IEEE Trans. Image Processing, 19(12): pages 3204-3221, décembre 2010. Mots-clés : Clustering algorithms, Image color analysis, Image edge detection, Image segmentation, Monte Carlo Sampling, Object-based stochastic model. @ARTICLE{mallet_tip2010, author = {Mallet, C. and Lafarge, F. and Roux, M. and Soergel, U. and Bretar, F. and Heipke, C.}, title = {A Marked Point Process for Modeling Lidar Waveforms}, year = {2010}, month = {décembre}, journal = {IEEE Trans. Image Processing}, volume = {19}, number = {12}, pages = {3204-3221}, url = {http://dx.doi.org/10.1109/TIP.2010.2052825}, keyword = {Clustering algorithms, Image color analysis, Image edge detection, Image segmentation, Monte Carlo Sampling, Object-based stochastic model} } Abstract : Lidar waveforms are 1D signals representing a train of echoes caused by reflections at different targets. Modeling these echoes with the appropriate parametric function is useful to retrieve information about the physical characteristics of the targets. This paper presents a new probabilistic model based on a marked point process which reconstructs the echoes from recorded discrete waveforms as a sequence of parametric curves. Such an approach allows to fit each mode of a waveform with the most suitable function and to deal with both, symmetric and asymmetric, echoes. The model takes into account a data term, which measures the coherence between the models and the waveforms, and a regularization term, which introduces prior knowledge on the reconstructed signal. The exploration of the associated configuration space is performed by a Reversible Jump Markov Chain Monte Carlo sampler coupled with simulated annealing. Experiments with different kinds of lidar signals, especially from urban scenes, show the high potential of the proposed approach. To further demonstrate the advantages of the suggested method, actual laser scans are classified and the results are reported.

2 - Estimation of Markov Random Field prior parameters using Markov chain Monte Carlo Maximum Likelihood. X. Descombes et R. Morris et J. Zerubia et M. Berthod. IEEE Trans. Image Processing, 8(7): pages 954-963, juillet 1999. Mots-clés : Markov processes, Monte Carlo methods, Potts model, Image segmentation, Maximum likelihood estimation . @ARTICLE{xd99c, author = {Descombes, X. and Morris, R. and Zerubia, J. and Berthod, M.}, title = {Estimation of Markov Random Field prior parameters using Markov chain Monte Carlo Maximum Likelihood}, year = {1999}, month = {juillet}, journal = {IEEE Trans. Image Processing}, volume = {8}, number = {7}, pages = {954-963}, pdf = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?isnumber=16772&arnumber=772239&count=14&index=6}, keyword = {Markov processes, Monte Carlo methods, Potts model, Image segmentation, Maximum likelihood estimation } } Abstract : Developments in statistics now allow maximum likelihood estimators for the parameters of Markov random fields (MRFs) to be constructed. We detail the theory required, and present an algorithm that is easily implemented and practical in terms of computation time. We demonstrate this algorithm on three MRF models-the standard Potts model, an inhomogeneous variation of the Potts model, and a long-range interaction model, better adapted to modeling real-world images. We estimate the parameters from a synthetic and a real image, and then resynthesize the models to demonstrate which features of the image have been captured by the model. Segmentations are computed based on the estimated parameters and conclusions drawn.

haut de la page

Article de conférence

1 - Fully Bayesian image segmentation-an engineering perspective. R. Morris et X. Descombes et J. Zerubia. Dans Proc. IEEE International Conference on Image Processing (ICIP), Vol. 3, pages 54-57, Santa Barbara, CA, USA, octobre 1997. Mots-clés : Bayes methods, Markov processes, Monte Carlo methods, Image sampling, Image segmentation. @INPROCEEDINGS{MorrisICIP97, author = {Morris, R. and Descombes, X. and Zerubia, J.}, title = { Fully Bayesian image segmentation-an engineering perspective}, year = {1997}, month = {octobre}, booktitle = {Proc. IEEE International Conference on Image Processing (ICIP)}, volume = {3}, pages = {54-57}, address = {Santa Barbara, CA, USA}, pdf = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=631978&isnumber=13718}, keyword = {Bayes methods, Markov processes, Monte Carlo methods, Image sampling, Image segmentation} } Abstract : Developments in Markov chain Monte Carlo procedures have made it possible to perform fully Bayesian image segmentation. By this we mean that all the parameters are treated identically, be they the segmentation labels, the class parameters or the Markov random field prior parameters. We perform the analysis by sampling from the posterior distribution of all the parameters. Sampling from the MRF parameters has traditionally been considered if not intractable then at least computationally prohibitive. In the statistics literature there are descriptions of experiments showing that the MRF parameters may be sampled by approximating the partition function. These experiments are all, however, on `toy' problems; for the typical size of image encountered in engineering applications the phase transition behaviour of the models becomes a major limiting factor in the estimation of the partition function. Nevertheless, we show that, with some care, fully Bayesian segmentation can be performed on realistic sized images. We also compare the fully Bayesian approach with the approximate pseudolikelihood method

haut de la page

affichage : 10 par page | 20 par page | 50 par page

Ces pages sont générées par