Karim Kalti

Image segmentation by adaptive distance based on EM algorithm

arXiv (Cornell University), Apr 7, 2012

This paper introduces a Bayesian image segmentation algorithm based on finite mixtures. An EM alg... more This paper introduces a Bayesian image segmentation algorithm based on finite mixtures. An EM algorithm is developed to estimate parameters of the Gaussian mixtures. The finite mixture is a flexible and powerful probabilistic modeling tool. It can be used to provide a model-based clustering in the field of pattern recognition. However, the application of finite mixtures to image segmentation presents some difficulties; especially it's sensible to noise. In this paper we propose a variant of this method which aims to resolve this problem. Our approach proceeds by the characterization of pixels by two features: the first one describes the intrinsic properties of the pixel and the second characterizes the neighborhood of pixel. Then the classification is made on the base on adaptive distance which privileges the one or the other features according to the spatial position of the pixel in the image. The obtained results have shown a significant improvement of our approach compared to the standard version of EM algorithm.

Download

Un système informatique d’archivage et d’indexation de cas d’IRM de tumeurs cérébrales

Feuillets de Radiologie, 2011

Objective. We describe the design and implementation of a computer system for archiving and index... more Objective. We describe the design and implementation of a computer system for archiving and indexing brain tumors as well as the creation of a database for MRI cases of six histological types. Material and methods. We explored the techniques offered by the incremental development of computer technology useful for the medical decision making process. We began by a remote server data archiving system, which we then implemented for a local indexing application. Results. The reasoning behind our approach, based on Bayesian networks, demonstrated good efficiency in processing indexing problems. Our main contribution is to modelize the radiologist's reasoning when interpreting medical cases. Real case testing argued in favor of our choice. Conclusion. Our model can be used as the core of an approved diagnosis aid system. The promising perspectives offered by our model could be useful in other approaches in the area of artificial intelligence.

Download

A Robust Contextual Fuzzy C-Means Clustering Algorithm for Noisy Image Segmentation

Journal of Classification

Special Issue on Image Processing and Analysis

19 | P a g e www.ijacsa.thesai.org Image segmentation by adaptive distance based on EM algorithm

A New Collaborative Classification Process for Microcalcification Detection Based on Graphs and Knowledge Propagation

Journal of Digital Imaging

$Research paper thumbnail of Etude de Mod\\eles\\a base de r\\'eseaux Bay\\'esiens pour l'aide au diagnostic de tumeurs c\\'er\\'ebrales$

Etude de Mod\\eles\\a base de r\\'eseaux Bay\\'esiens pour l'aide au diagnostic de tumeurs c\\'er\\'ebrales

arXiv preprint arXiv:1207.2459, 2012

Abstract: This article describes different models based on Bayesian networks RB modeling expertis... more

Download

Image Segmentation Based on Adaptive Fuzzy-C-Means Clustering

2010 20th International Conference on Pattern Recognition, 2010

... [6] DL Pham. Spatial models for fuzzy clustering, Computer vision and image understanding, Vo... more

A probabilistic network based similiarity measure for cerebral tumors MRI cases retrieval

by Hedi YAZID, Karim Kalti, and Najoua Essoukri Ben Amara

2011 IEEE Third International Workshop On Computational Intelligence In Medical Imaging, 2011

... For Cerebral Tumors MRI Cases Retrieval HediYAZID, Karim KALTI, Najoua BEN AMARA ESSOUKRI Nat... more

MRI cases containing cerebral tumors retrieval using Bayesian networks

by Hedi YAZID and Karim Kalti

The 10th IEEE International Symposium on Signal Processing and Information Technology, 2010

... TUu2 1 Engineers National school Sousse, Tunisia e-mail: hedi.yazid@gmail.com.karim.kalti@ gm... more

Image segmentation by adaptive distance based on EM algorithm

International Journal of Advanced Computer Science and Applications, 2011

This paper introduces a Bayesian image segmentation algorithm based on finite mixtures. An EM alg... more This paper introduces a Bayesian image segmentation algorithm based on finite mixtures. An EM algorithm is developed to estimate parameters of the Gaussian mixtures. The finite mixture is a flexible and powerful probabilistic modeling tool. It can be used to provide a model-based clustering in the field of pattern recognition. However, the application of finite mixtures to image segmentation presents some difficulties; especially it's sensible to noise. In this paper we propose a variant of this method which aims to resolve this problem. Our approach proceeds by the characterization of pixels by two features: the first one describes the intrinsic properties of the pixel and the second characterizes the neighborhood of pixel. Then the classification is made on the base on adaptive distance which privileges the one or the other features according to the spatial position of the pixel in the image. The obtained results have shown a significant improvement of our approach compared to the standard version of EM algorithm.

Download

Software Comparison Dealing with Bayesian Networks

Advances in Neural Networks – ISNN 2011, 2011

... format). SAMIAM Samiam has two main components: a graphical interface and an inference engine... more

Image Segmentation by Gaussian Mixture Models and Modified FCM Algorithm

Abstract: The Expectation Maximization (EM) algorithm and the clustering method Fuzzy-C-Means (FC... more Abstract: The Expectation Maximization (EM) algorithm and the clustering method Fuzzy-C-Means (FCM) are widely used in image segmentation. However, the major drawback of these methods is their sensitivity to the noise. In this paper, we propose a variant of these methods which aim at resolving this problem. Our approaches proceed by the characterization of pixels by two features: the first one describes the intrinsic properties of the pixel and the second characterizes the neighborhood of pixel. Then, the classification is made on the base on adaptive distance which privileges the one or the other features according to the spatial position of the pixel in the image. The obtained results have shown a significant improvement of our approaches performance compared to the standard version of the EM and FCM, respectively, especially regarding about the robustness face to noise and the accuracy of the edges between regions.

Download

A New Labeled Quadtree-based Distance for Medical Image Retrieval

Image segmentation by gaussian mixture models and modified FCM algorithm

Int. Arab J. Inf. Technol., 2014

The Expectation Maximization (EM) algorithm and the clustering method Fuzzy(C(Means (FCM) are wid... more The Expectation Maximization (EM) algorithm and the clustering method Fuzzy(C(Means (FCM) are widely used in image segmentation. However, the major drawback of these methods is their sensitivity to the noise. In this paper, we propose a variant of these methods which aim at resolving this problem. Our approaches proceed by the characterization of pixels by two features: the first one describes the intrinsic properties of the pixel and the second characterizes the neighborhood of pixel. Then, the classification is made on the base on adaptive distance which privileges the one or the other features according to the spatial position of the pixel in the image. The obtained results have shown a significant improvement of our approaches performance compared to the standard version of the EM and FCM, respectively, especially regarding about the robustness face to noise and the accuracy of the edges between regions.

Superpixel Based Segmentation of Historical Document Images Using a Multiscale Texture Analysis

In this paper, a superpixel based segmentation of Historical Document Images (HDIs) using multisc... more In this paper, a superpixel based segmentation of Historical Document Images (HDIs) using multiscale texture analysis is proposed. A Simple Linear Iterative Clustering (SLIC) superpixel technique and Kmeans classifier are applied in order to separate the input image into background and foreground superpixels. The foreground superpixels are characterized by the standard deviation and the mean of the Gabor features. These features are extracted in a multiscale fashion to adapt to the variability of the textures that may be present in HDIs. Text/graphic separation is then performed by applying a classification of the foreground superpixels for each texture analysis scale followed by a merging step of the obtained classification results. Since the classification results depend on the used classifier, a comparative study is performed for supervised (Support Vector Machine (SVM), K-Nearest Neighbors (KNN)) and unsupervised (Kmeans, Fuzzy C-Means (FCM)) techniques. Experiments show the eff...

A new conditional region growing approach for an accurate detection of microcalcifications from mammographic images

In this paper, we propose a new Conditional Region Growing (CRG) approach with the ability of fin... more In this paper, we propose a new Conditional Region Growing (CRG) approach with the ability of finding the accurate MC boundaries starting from selected seed points. The starting seed points are determined based on regional maxima detection and superpixel analysis. The region growing step is controlled by a set of criteria derived from prior knowledge to characterize MCs. The key feature is to highlight below each MC to estimate the appropriate criteria and not to use the same parameters for all of them. Defined criteria can be divided into two categories. The first one concerns the neighbourhood searching size. The second one deals with the gradient information and shape evolution within the growing process. Experimental results show the benefits of used criteria in terms of improving the MC delineation qualities.

Morphological-based microcalcification detection using adaptive thresholding and structural similarity indices

In this paper, we propose a new morphological-based method for automatic detection of microcalcif... more In this paper, we propose a new morphological-based method for automatic detection of microcalcifications in digitized mammograms. It uses various structuring elements to deal with the diversity of microcalcification characteristics. The obtained morphological maps are converted to a continuous suspicion map (SM) based on the structural similarity index (SSIM). This new semantic representation map is then locally analyzed, using superpixels, to automatically estimate adaptive threshold values and finally identify potential microcalcification areas. The proposed method was evaluated using the publicly-available INBreast database. Experimental results show the benefits gained in terms of improving microcalcification detection performances compared to some state-of-the-art methods.

Iterative Fuzzy Segmentation for Breast Skin-line Detection

Image segmentation using MRF model optimized by a hybrid ACO-ICM algorithm

Microcalcifications detection from mammographic images based on region growing and variational energy convergence

2018 4th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)

In this paper we present a new region growing approach for microcalcification detection in mammog... more In this paper we present a new region growing approach for microcalcification detection in mammographie images. This approach is based on combining a variational approach concept and the Region Growing algorithm in order to correctly delineate the microcalcifications (MCs) that appear in a Region Of Interest (ROI). Its main goal is to make an initial seeds evolve towards a meaningful partition of the ROI, using the minimization of such a region-based energy criterion. The obtained results show encouraging results in terms of reduction of false negative detections and in terms of the delineation of the targeted microcalcifications from the ROIs.

Uploads

Papers by Karim Kalti

Log In