High-quality region-based foreground segmentation using a spatial grid of SVM classifiers

This paper presents a novel background modeling system that uses a spatial grid of Support Vector Machines classifiers for segmenting moving objects, which is a key step in many video-based consumer applications. The system is able to adapt to a large range of dynamic background situations since no parametric model or statistical distribution are assumed. This is achieved by using a different classifier per image region that learns the specific appearance of that scene region and its variations (illumination changes, dynamic backgrounds, etc.). The proposed system has been tested with a recent public database, outperforming other state-of-the-art algorithms.

Advanced background modeling with RGB-D sensors through classifiers combination and inter-frame foreground prediction

An innovative background modeling technique that is able to accurately segment foreground regions in RGB-D imagery (RGB plus depth) has been presented in this paper. The technique is based on a Bayesian framework that efficiently fuses different sources of information to segment the foreground. In particular, the final segmentation is obtained by considering a prediction of the foreground regions, carried out by a novel Bayesian Network with a depth-based dynamic model, and, by considering two independent depth and color-based mixture of Gaussians background models. The efficient Bayesian combination of all these data reduces the noise and uncertainties introduced by the color and depth features and the corresponding models. As a result, more compact segmentations, and refined foreground object silhouettes are obtained. Experimental results with different databases suggest that the proposed technique outperforms existing state-of-the-art algorithms.

Background foreground segmentation with RGB-D Kinect data: An efficient combination of classifiers

Low cost RGB-D cameras such as the Microsoft’s Kinect or the Asus’s Xtion Pro are completely changing the computer vision world, as they are being successfully used in several applications and research areas. Depth data are particularly attractive and suitable for applications based on moving objects detection through foreground/background segmentation approaches; the RGB-D applications proposed in literature employ, in general, state of the art foreground/background segmentation techniques based on the depth information without taking into account the color information. The novel approach that we propose is based on a combination of classifiers that allows improving background subtraction accuracy with respect to state of the art algorithms by jointly considering color and depth data. In particular, the combination of classifiers is based on a weighted average that allows to adaptively modifying the support of each classifier in the ensemble by considering foreground detections in the previous frames and the depth and color edges. In this way, it is possible to reduce false detections due to critical issues that can not be tackled by the individual classifiers such as: shadows and illumination changes, color and depth camouflage, moved background objects and noisy depth measurements. Moreover, we propose, for the best of the author’s knowledge, the first publicly available RGB-D benchmark dataset with hand-labeled ground truth of several challenging scenarios to test background/foreground segmentation algorithms.

Decision boundary for discrete Bayesian network classifiers

Bayesian network classifiers are a powerful machine learning tool. In order to evaluate the expressive power of these models, we compute families of polynomials that sign-represent decision functions induced by Bayesian network classifiers. We prove that those families are linear combinations of products of Lagrange basis polynomials. In absence of V -structures in the predictor sub-graph, we are also able to prove that this family of polynomials does indeed characterize the specific classifier considered. We then use this representation to bound the number of decision functions representable by Bayesian network classifiers with a given structure.

Oxytocin releases atrial natriuretic peptide by combining with oxytocin receptors in the heart

Previous studies indicated that the central nervous system induces release of the cardiac hormone atrial natriuretic peptide (ANP) by release of oxytocin from the neurohypophysis. The presence of specific transcripts for the oxytocin receptor was demonstrated in all chambers of the heart by amplification of cDNA by the PCR using specific oligonucleotide primers. Oxytocin receptor mRNA content in the heart is 10 times lower than in the uterus of female rats. Oxytocin receptor transcripts were demonstrated by in situ hybridization in atrial and ventricular sections and confirmed by competitive binding assay using frozen heart sections. Perfusion of female rat hearts for 25 min with Krebs–Henseleit buffer resulted in nearly constant release of ANP. Addition of oxytocin (10−6 M) significantly stimulated ANP release, and an oxytocin receptor antagonist (10−7 and 10−6 M) caused dose-related inhibition of oxytocin-induced ANP release and in the last few minutes of perfusion decreased ANP release below that in control hearts, suggesting that intracardiac oxytocin stimulates ANP release. In contrast, brain natriuretic peptide release was unaltered by oxytocin. During perfusion, heart rate decreased gradually and it was further decreased significantly by oxytocin (10−6 M). This decrease was totally reversed by the oxytocin antagonist (10−6 M) indicating that oxytocin released ANP that directly slowed the heart, probably by release of cyclic GMP. The results indicate that oxytocin receptors mediate the action of oxytocin to release ANP, which slows the heart and reduces its force of contraction to produce a rapid reduction in circulating blood volume.

Potent inhibition of tumor survival in vivo by β-lapachone plus taxol: Combining drugs imposes different artificial checkpoints

Ablation of tumor colonies was seen in a wide spectrum of human carcinoma cells in culture after treatment with the combination of β-lapachone and taxol, two low molecular mass compounds. They synergistically induced death of cultured ovarian, breast, prostate, melanoma, lung, colon, and pancreatic cancer cells. This synergism is schedule dependent; namely, taxol must be added either simultaneously or after β-lapachone. This combination therapy has unusually potent antitumor activity against human ovarian and prostate tumor prexenografted in mice. There is little host toxicity. Cells can commit to apoptosis at cell-cycle checkpoints, a mechanism that eliminates defective cells to ensure the integrity of the genome. We hypothesize that when cells are treated simultaneously with drugs activating more than one different cell-cycle checkpoint, the production of conflicting regulatory signaling molecules induces apoptosis in cancer cells. β-Lapachone causes cell-cycle delays in late G1 and S phase, and taxol arrests cells at G2/M. Cells treated with both drugs were delayed at multiple checkpoints before committing to apoptosis. Our findings suggest an avenue for developing anticancer therapy by exploiting apoptosis-prone “collisions” at cell-cycle checkpoints.

A specific transition state for S-peptide combining with folded S-protein and then refolding

We measured the folding and unfolding kinetics of mutants for a simple protein folding reaction to characterize the structure of the transition state. Fluorescently labeled S-peptide analogues combine with S-protein to form ribonuclease S analogues: initially, S-peptide is disordered whereas S-protein is folded. The fluorescent probe provides a convenient spectroscopic probe for the reaction. The association rate constant, kon, and the dissociation rate constant, koff, were both determined for two sets of mutants. The dissociation rate constant is measured by adding an excess of unlabeled S-peptide analogue to a labeled complex (RNaseS*). This strategy allows kon and koff to be measured under identical conditions so that microscopic reversibility applies and the transition state is the same for unfolding and refolding. The first set of mutants tests the role of the α-helix in the transition state. Solvent-exposed residues Ala-6 and Gln-11 in the α-helix of native RNaseS were replaced by the helix destabilizing residues glycine or proline. A plot of log kon vs. log Kd for this series of mutants is linear over a very wide range, with a slope of −0.3, indicating that almost all of the molecules fold via a transition state involving the helix. A second set of mutants tests the role of side chains in the transition state. Three side chains were investigated: Phe-8, His-12, and Met-13, which are known to be important for binding S-peptide to S-protein and which also contribute strongly to the stability of RNaseS*. Only the side chain of Phe-8 contributes significantly, however, to the stability of the transition state. The results provide a remarkably clear description of a folding transition state.

Inactivation of the mouse sperm receptor, mZP3, by site-directed mutagenesis of individual serine residues located at the combining site for sperm

To initiate fertilization, mouse sperm bind to Ser- (O-) linked oligosaccharides located at the sperm combining site of zona pellucida glycoprotein mZP3. Apparently, the oligosaccharides are present on one or more of five Ser residues clustered in the carboxyl-terminal region of the mZP3 polypeptide. Here, each of the Ser residues, as well as an intervening Asn residue, was converted to a small, nonhydroxy amino acid by site-directed mutagenesis. Mouse embryonal carcinoma (EC) cells were then stably transfected with the wild-type and mutated mZP3 genes. In each case, transfected cells synthesized and secreted recombinant EC-mZP3 into the culture medium. The glycoproteins were partially purified and assayed for their ability to inhibit binding of sperm to ovulated eggs in vitro. As compared with wild-type EC-mZP3, mutations of Ser-329, Ser-331, or Ser-333 had no effect on sperm receptor activity. Mutation of Asn-330, a potential N-linked glycosylation site, also had no effect on sperm receptor activity. On the other hand, mutation of either Ser-332 or Ser-334, or mutation of Ser-332, Ser-333, and Ser-334, resulted in complete inactivation of EC-mZP3 as a sperm receptor. These results suggest that Ser-332 and Ser-334, residues conserved in mouse, hamster, and human ZP3, are essential for sperm receptor activity.

Nueva técnica de fusión de clasificadores aplicada a la mejora de la segmentación de audio

Este artículo presenta un nuevo algoritmo de fusión de clasificadores a partir de su matriz de confusión de la que se extraen los valores de precisión (precision) y cobertura (recall) de cada uno de ellos. Los únicos datos requeridos para poder aplicar este nuevo método de fusión son las clases o etiquetas asignadas por cada uno de los sistemas y las clases de referencia en la parte de desarrollo de la base de datos. Se describe el algoritmo propuesto y se recogen los resultados obtenidos en la combinación de las salidas de dos sistemas participantes en la campaña de evaluación de segmentación de audio Albayzin 2012. Se ha comprobado la robustez del algoritmo, obteniendo una reducción relativa del error de segmentación del 6.28% utilizando para realizar la fusión el sistema con menor y mayor tasa de error de los presentados a la evaluación.

Combining visual features and Growing Neural Gas networks for robotic 3D SLAM

The use of 3D data in mobile robotics provides valuable information about the robot’s environment. Traditionally, stereo cameras have been used as a low-cost 3D sensor. However, the lack of precision and texture for some surfaces suggests that the use of other 3D sensors could be more suitable. In this work, we examine the use of two sensors: an infrared SR4000 and a Kinect camera. We use a combination of 3D data obtained by these cameras, along with features obtained from 2D images acquired from these cameras, using a Growing Neural Gas (GNG) network applied to the 3D data. The goal is to obtain a robust egomotion technique. The GNG network is used to reduce the camera error. To calculate the egomotion, we test two methods for 3D registration. One is based on an iterative closest points algorithm, and the other employs random sample consensus. Finally, a simultaneous localization and mapping method is applied to the complete sequence to reduce the global error. The error from each sensor and the mapping results from the proposed method are examined.

Improving classification using a Confidence Matrix based on weak classifiers applied to OCR

This paper proposes a new feature representation method based on the construction of a Confidence Matrix (CM). This representation consists of posterior probability values provided by several weak classifiers, each one trained and used in different sets of features from the original sample. The CM allows the final classifier to abstract itself from discovering underlying groups of features. In this work the CM is applied to isolated character image recognition, for which several set of features can be extracted from each sample. Experimentation has shown that the use of CM permits a significant improvement in accuracy in most cases, while the others remain the same. The results were obtained after experimenting with four well-known corpora, using evolved meta-classifiers with the k-Nearest Neighbor rule as a weak classifier and by applying statistical significance tests.

Optimal Design of a Hybrid Membrane System Combining Reverse and Forward Osmosis for Seawater Desalination

In this work we study Forward Osmosis (FO) as an emerging desalination technology, and its capability to replace totally or partially Reverse Osmosis (RO) in order to reduce the great amount of energy required in the current desalination plants. For this purpose, we propose a superstructure that includes both membrane based desalination technologies, allowing the selection of only one of the technologies or a combination of both of them seeking for the optimal configuration of the network. The optimization problem is solved for a seawater desalination plant with a given fresh water production. The results obtained show that the optimal solution combines both desalination technologies to reduce not only the energy consumption but also the total cost of the desalination process in comparison with the same plant but operating only with RO.

Mechanisms that Trigger a Good Health-Care Response to Intimate Partner Violence in Spain. Combining Realist Evaluation and Qualitative Comparative Analysis Approaches

Background. Health care professionals, especially those working in primary health-care services, can play a key role in preventing and responding to intimate partner violence. However, there are huge variations in the way health care professionals and primary health care teams respond to intimate partner violence. In this study we tested a previously developed programme theory on 15 primary health care center teams located in four different Spanish regions: Murcia, C Valenciana, Castilla-León and Cantabria. The aim was to identify the key combinations of contextual factors and mechanisms that trigger a good primary health care center team response to intimate partner violence. Methods. A multiple case-study design was used. Qualitative and quantitative information was collected from each of the 15 centers (cases). In order to handle the large amount of information without losing familiarity with each case, qualitative comparative analysis was undertaken. Conditions (context and mechanisms) and outcomes, were identified and assessed for each of the 15 cases, and solution formulae were calculated using qualitative comparative analysis software. Results. The emerging programme theory highlighted the importance of the combination of each team’s self-efficacy, perceived preparation and women-centredness in generating a good team response to intimate partner violence. The use of the protocol and accumulated experience in primary health care were the most relevant contextual/intervention conditions to trigger a good response. However in order to achieve this, they must be combined with other conditions, such as an enabling team climate, having a champion social worker and having staff with training in intimate partner violence. Conclusions. Interventions to improve primary health care teams’ response to intimate partner violence should focus on strengthening team’s self-efficacy, perceived preparation and the implementation of a woman-centred approach. The use of the protocol combined with a large working experience in primary health care, and other factors such as training, a good team climate, and having a champion social worker on the team, also played a key role. Measures to sustain such interventions and promote these contextual factors should be encouraged.