Improving Clustering Algorithms for Image Segmentation using Contour and Region Information

In image segmentation, clustering algorithms are very popular because they are intuitive and, some of them, easy to implement. For instance, the k-means is one of the most used in the literature, and many authors successfully compare their new proposal with the results achieved by the k-means. However, it is well known that clustering image segmentation has many problems. For instance, the number of regions of the image has to be known a priori, as well as different initial seed placement (initial clusters) could produce different segmentation results. Most of these algorithms could be slightly improved by considering the coordinates of the image as features in the clustering process (to take spatial region information into account). In this paper we propose a significant improvement of clustering algorithms for image segmentation. The method is qualitatively and quantitative evaluated over a set of synthetic and real images, and compared with classical clustering approaches. Results demonstrate the validity of this new approach

Is the ability of biocontrol fluorescent pseudomonads to produce the antifungal metabolite 2,4-diacetylphloroglucinol really synonymous with higher plant protection?

The antifungal compound 2,4-diacetylphloroglucinol (Phl) contributes to biocontrol in pseudomonads, but whether or not Phl(+) biocontrol pseudomonads display higher plant-protecting activity than Phl(-) biocontrol pseudomonads remains to be demonstrated. This issue was addressed by assessing 230 biocontrol fluorescent pseudomonads selected from a collection of 3132 bacterial isolates obtained from 63 soils worldwide. One-third of the biocontrol pseudomonads were Phl(+) and almost all Phl(+) isolates also produced hydrogen cyanide (HCN). The only Phl(+) HCN(-) strain did harbor hcn genes, but with the deletion of a 134 bp hcnC fragment corresponding to an ADP-binding motif. Statistical analysis of biocontrol isolate distributions indicated that Phl production ability was associated with superior disease suppression activity in the Pythium-cucumber and Fusarium-tomato pathosystems, but this was also the case with HCN production ability. However, HCN significance was not as strong, as indicated both by the comparison of Phl(-) HCN(+) and Phl(-) HCN(-) strains and by correlation analyses. This is the first population-level demonstration of the higher plant-protecting activity of Phl(+) biocontrol pseudomonads in comparison with Phl(-) biocontrol pseudomonads.

An information theoretic framework for image segmentation

In this paper, an information theoretic framework for image segmentation is presented. This approach is based on the information channel that goes from the image intensity histogram to the regions of the partitioned image. It allows us to define a new family of segmentation methods which maximize the mutual information of the channel. Firstly, a greedy top-down algorithm which partitions an image into homogeneous regions is introduced. Secondly, a histogram quantization algorithm which clusters color bins in a greedy bottom-up way is defined. Finally, the resulting regions in the partitioning algorithm can optionally be merged using the quantized histogram

Segmentation automatique de la matière grise et de la matière blanche par un algorithme génétique : une étude de validation

Résumé Objectif: l'observation des variations de volume de la matière grise (MG), de la matière blanche (MB), et du liquide céphalo-rachidien (LCR) est particulièrement utile dans l'étude de nombreux processus physiopathologiques, la mesure quantitative 'in vivo' de ces volumes présente un intérêt considérable tant en recherche qu'en pratique clinique. Cette étude présente et valide une méthode de segmentation automatique du cerveau avec mesure des volumes de MG et MB sur des images de résonance magnétique. Matériel et Méthode: nous utilisons un algorithme génétique automatique pour segmenter le cerveau en MG, MB et LCR à partir d'images tri-dimensionnelles de résonance magnétique en pondération Ti. Une étude morphométrique a été conduite sur 136 sujets hommes et femmes de 15 à 74 ans. L'algorithme a ensuite été validé par 5 approches différentes: I. Comparaison de mesures de volume sur un cerveau de cadavre par méthode automatique et par mesure de déplacement d'eau selon la méthode d'Archimède. 2. Comparaison de mesures surfaces sur des images bidimensionnelles segmentées soit par un traçage manuel soit par la méthode automatique. 3. Evaluation de la fiabilité de la segmentation par acquisitions et segmentations itératives du même cerveau. 4. Les volumes de MG, MB et LCR ont été utilisés pour une étude du vieillissement normal de la population. 5. Comparaison avec les données existantes de la littérature. Résultats: nous avons pu observer une variation de la mesure de 4.17% supplémentaire entre le volume d'un cerveau de cadavre mesuré par la méthode d'Archimède, en majeure partie due à la persistance de tissus après dissection_ La comparaison des méthodes de comptage manuel de surface avec la méthode automatique n'a pas montré de variation significative. L'épreuve du repositionnement du même sujet à diverses reprises montre une très bonne fiabilité avec une déviation standard de 0.46% pour la MG, 1.02% pour la MB et 3.59% pour le LCR, soit 0.19% pour le volume intracrânien total (VICT). L'étude morphométrique corrobore les résultats des études anatomiques et radiologiques existantes. Conclusion: la segmentation du cerveau par un algorithme génétique permet une mesure 100% automatique, fiable et rapide des volumes cérébraux in vivo chez l'individu normal.

The ability of haemolysins expressed by atypical enteropathogenic Escherichia coli to bind to extracellular matrix components

Typical and atypical enteropathogenic Escherichia coli (EPEC) are considered important bacterial causes of diarrhoea. Considering the repertoire of virulence genes, atypical EPEC (aEPEC) is a heterogeneous group, harbouring genes that are found in other diarrheagenic E. coli pathotypes, such as those encoding haemolysins. Haemolysins are cytolytic toxins that lyse host cells disrupting the function of the plasma membrane. In addition, these cytolysins mediate a connection to vascular tissue and/or blood components, such as plasma and cellular fibronectin. Therefore, we investigated the haemolytic activity of 72 aEPEC isolates and determined the correlation of this phenotype with the presence of genes encoding enterohaemolysins (Ehly) and cytolysin A (ClyA). In addition, the correlation between the expression of haemolysins and the ability of these secreted proteins to adhere to extracellular matrix (ECM) components was also assessed in this study. Our findings demonstrate that a subset of aEPEC presents haemolytic activity due to the expression of Ehlys and/or ClyA and that this activity is closely related to the ability of these isolates to bind to ECM components.

Predictive ability of an early diagnostic guess in patients presenting with chest pain; a longitudinal descriptive study.

ABSTRACT: BACKGROUND: The intuitive early diagnostic guess could play an important role in reaching a final diagnosis. However, no study to date has attempted to quantify the importance of general practitioners' (GPs) ability to correctly appraise the origin of chest pain within the first minutes of an encounter. METHODS: The validation study was nested in a multicentre cohort study with a one year follow-up and included 626 successive patients who presented with chest pain and were attended by 58 GPs in Western Switzerland. The early diagnostic guess was assessed prior to a patient's history being taken by a GP and was then compared to a diagnosis of chest pain observed over the next year. RESULTS: Using summary measures clustered at the GP's level, the early diagnostic guess was confirmed by further investigation in 51.0% (CI 95%; 49.4% to 52.5%) of patients presenting with chest pain. The early diagnostic guess was more accurate in patients with a life threatening illness (65.4%; CI 95% 64.5% to 66.3%) and in patients who did not feel anxious (62.9%; CI 95% 62.5% to 63.3%). The predictive abilities of an early diagnostic guess were consistent among GPs. CONCLUSIONS: The GPs early diagnostic guess was correct in one out of two patients presenting with chest pain. The probability of a correct guess was higher in patients with a life-threatening illness and in patients not feeling anxious about their pain.

Multimodal evaluation for medical image segmentation

This paper is a joint effort between five institutionsthat introduces several novel similarity measures andcombines them to carry out a multimodal segmentationevaluation. The new similarity measures proposed arebased on the location and the intensity values of themisclassified voxels as well as on the connectivity andthe boundaries of the segmented data. We showexperimentally that the combination of these measuresimprove the quality of the evaluation. The study that weshow here has been carried out using four differentsegmentation methods from four different labs applied toa MRI simulated dataset of the brain. We claim that ournew measures improve the robustness of the evaluation andprovides better understanding about the differencebetween segmentation methods.

Epithelial Na+ channel mutants causing Liddle's syndrome retain ability to respond to aldosterone and vasopressin.

Liddle's syndrome is a monogenic form of hypertension caused by mutations in the PY motif of the COOH terminus of beta- and gamma-epithelial Na+ channel (ENaC) subunits. These mutations lead to retention of active channels at the cell surface. Because of the critical role of this PY motif in the stability of ENaCs at the cell surface, we have investigated its contribution to the ENaC response to aldosterone and vasopressin. Mutants of the PY motif in beta- and gamma-ENaC subunits (beta-Y618A, beta-P616L, beta-R564stop, and gamma-K570stop) were stably expressed by retroviral gene transfer in a renal cortical collecting duct cell line (mpkCCDcl4), and transepithelial Na+ transport was assessed by measurements of the benzamil-sensitive short-circuit current (Isc). Cells that express ENaC mutants of the PY motif showed a five- to sixfold higher basal Isc compared with control cells and responded to stimulation by aldosterone (10(-6) M) or vasopressin (10(-9) M) with a further increase in Isc. The rates of the initial increases in Isc after aldosterone or vasopressin stimulation were comparable in cells transduced with wild-type and mutant ENaCs, but reversal of the effects of aldosterone and vasopressin was slower in cells that expressed the ENaC mutants. The conserved sensitivity of ENaC mutants to stimulation by aldosterone and vasopressin together with the prolonged activity at the cell surface likely contribute to the increased Na+ absorption in the distal nephron of patients with Liddle's syndrome.

3D Automated Lung Nodule Segmentation in HRCT

A fully-automated 3D image analysis method is proposed to segment lung nodules in HRCT. A specific gray-level mathematical morphology operator, the SMDC-connection cost, acting in the 3D space of the thorax volume is defined in order to discriminate lung nodules from other dense (vascular) structures. Applied to clinical data concerning patients with pulmonary carcinoma, the proposed method detects isolated, juxtavascular and peripheral nodules with sizes ranging from 2 to 20 mm diameter. The segmentation accuracy was objectively evaluated on real and simulated nodules. The method showed a sensitivity and a specificity ranging from 85% to 97% and from 90% to 98%, respectively.

Rank estimation of trajectory matrix in motion segmentation

A novel technique for estimating the rank of the trajectory matrix in the local subspace affinity (LSA) motion segmentation framework is presented. This new rank estimation is based on the relationship between the estimated rank of the trajectory matrix and the affinity matrix built with LSA. The result is an enhanced model selection technique for trajectory matrix rank estimation by which it is possible to automate LSA, without requiring any a priori knowledge, and to improve the final segmentation

Enhanced Model Selection for motion segmentation

In this paper a novel rank estimation technique for trajectories motion segmentation within the Local Subspace Affinity (LSA) framework is presented. This technique, called Enhanced Model Selection (EMS), is based on the relationship between the estimated rank of the trajectory matrix and the affinity matrix built by LSA. The results on synthetic and real data show that without any a priori knowledge, EMS automatically provides an accurate and robust rank estimation, improving the accuracy of the final motion segmentation

Harmonic active contours.

We propose a segmentation method based on the geometric representation of images as 2-D manifolds embedded in a higher dimensional space. The segmentation is formulated as a minimization problem, where the contours are described by a level set function and the objective functional corresponds to the surface of the image manifold. In this geometric framework, both data-fidelity and regularity terms of the segmentation are represented by a single functional that intrinsically aligns the gradients of the level set function with the gradients of the image and results in a segmentation criterion that exploits the directional information of image gradients to overcome image inhomogeneities and fragmented contours. The proposed formulation combines this robust alignment of gradients with attractive properties of previous methods developed in the same geometric framework: 1) the natural coupling of image channels proposed for anisotropic diffusion and 2) the ability of subjective surfaces to detect weak edges and close fragmented boundaries. The potential of such a geometric approach lies in the general definition of Riemannian manifolds, which naturally generalizes existing segmentation methods (the geodesic active contours, the active contours without edges, and the robust edge integrator) to higher dimensional spaces, non-flat images, and feature spaces. Our experiments show that the proposed technique improves the segmentation of multi-channel images, images subject to inhomogeneities, and images characterized by geometric structures like ridges or valleys.