Spelling suggestions: "subject:"hidden markov model"" "subject:"hidden markov godel""
251 |
Détection non supervisée d'évènements rares dans un flot vidéo : application à la surveillance d'espaces publics / Unsupervised detection of rare events in a video stream : application to the surveillance of public spacesLuvison, Bertrand 13 December 2010 (has links)
Cette thèse est une collaboration entre le LAboratoire des Sciences et Matériaux pour l’Électronique et d’Automatique (LASMEA) de Clermont-Ferrand et le Laboratoire Vision et Ingénierie des Contenus (LVIC) du CEA LIST à Saclay. La première moitié de la thèse a été accomplie au sein de l’équipe ComSee (1) du LASMEA et la deuxième au LVIC. L’objectif de ces travaux est de concevoir un système de vidéo-assistance temps réel pour la détection d’évènements dans des scènes possiblement denses.La vidéosurveillance intelligente de scènes denses telles que des foules est particulièrement difficile, principalement à cause de leur complexité et de la grande quantité de données à traiter simultanément. Le but de cette thèse consiste à élaborer une méthode de détection d’évènements rares dans de telles scènes, observées depuis une caméra fixe. La méthode en question s’appuie sur l’analyse automatique de mouvement et ne nécessite aucune information à priori. Les mouvements nominaux sont déterminés grâce à un apprentissage statistique non supervisé. Les plus fréquemment observés sont considérés comme des évènements normaux. Une phase de classification permet ensuite de détecter les mouvements déviant trop du modèle statistique, pour les considérer comme anormaux. Cette approche est particulièrement adaptée aux lieux de déplacements structurés, tels que des scènes de couloirs ou de carrefours routiers. Aucune étape de calibration, de segmentation de l’image, de détection d’objets ou de suivi n’est nécessaire. Contrairement aux analyses de trajectoires d’objets suivis, le coût calculatoire de notre méthode est invariante au nombre de cibles présentes en même temps et fonctionne en temps réel. Notre système s’appuie sur une classification locale du mouvement de la scène, sans calibration préalable. Dans un premier temps, une caractérisation du mouvement est réalisée, soit par des méthodes classiques de flot optique, soit par des descripteurs spatio-temporels. Ainsi, nous proposons un nouveau descripteur spatio-temporel fondé sur la recherche d’une relation linéaire entre les gradients spatiaux et les gradients temporels en des zones où le mouvement est supposé uniforme. Tout comme les algorithmes de flot optique, ce descripteur s’appuie sur la contrainte d’illumination constante.Cependant en prenant en compte un voisinage temporel plus important, il permet une caractérisation du mouvement plus lisse et plus robuste au bruit. De plus, sa faible complexité calculatoire est bien adaptée aux applications temps réel. Nous proposons ensuite d’étudier différentes méthodes de classification : La première, statique, dans un traitement image par image, s’appuie sur une estimation bayésienne de la caractérisation du mouvement au travers d’une approche basée sur les fenêtres de Parzen. Cette nouvelle méthode est une variante parcimonieuse des fenêtres de Parzen. Nous montrons que cette approche est algorithmiquement efficace pour approximer de manière compacte et précise les densités de probabilité. La seconde méthode, basée sur les réseaux bayésiens, permet de modéliser la dynamique du mouvement. Au lieu de considérer ce dernier image par image, des séquences de mouvements sont analysées au travers de chaînes de Markov Cachées. Ajouté à cela, une autre contribution de ce manuscrit est de prendre en compte la modélisation du voisinage d’un bloc afin d’ajouter une cohérence spatiale à la propagation du mouvement. Ceci est réalisé par le biais de couplages de chaînes de Markov cachées.Ces différentes approches statistiques ont été évaluées sur des données synthétiques ainsi qu’en situations réelles, aussi bien pour la surveillance du trafic routier que pour la surveillance de foule.Cette phase d’évaluation permet de donner des premières conclusions encourageantes quant à la faisabilité de la vidéosurveillance intelligente d’espaces possiblement denses. / The automatic analysis of crowded areas in video sequences is particularly difficult because ofthe large amount of information to be processed simultaneously and the complexity of the scenes. We propose in this thesis a method for detecting abnormal events in possibly dense scenes observed from a static camera. The approach is based on the automatic classification of motion requiring no prior information. Motion patterns are encoded in an unsupervised learning framework in order to generate a statistical model of frequently observed (aka. normal) events. Then at the detection stage, motion patterns that deviate from the model are classified as unexpected events. The method is particularly adapted to scenes with structured movement with directional flow of objects or people such as corridors, roads, intersections. No camera calibration is needed, nor image segmentation, object detection and tracking. In contrast to approaches that rely on trajectory analysis of tracked objects, our method is independent of the number of targets and runs in real-time. Our system relies on a local classification of global scene movement. The local analysis is done on each blocks of a regular grid. We first introduce a new spatio-temporal local descriptor to characterize the movement efficiently. Assuming a locally uniform motion of space-time blocks of the image, our approach consists in determining whether there is a linear relationship between spatial gradients and temporal gradients. This spatio-temporal descriptor holds the Illumination constancy constraint like optical flow techniques, but it allows taking into account the spatial neighborhood and a temporal window by giving a smooth characterization of the motion, which makes it more robust to noise. In addition, its low computational complexity is suitable for real-time applications. Secondly, we present two different classification frameworks : The first approach is a static (frame by frame) classification approach based on a Bayesian characterization of the motion by using an approximation of the Parzen windowing method or Kernel Density Estimation (KDE) to model the probability density function of motion patterns.This new method is the sparse variant of the KDE (SKDE). We show that the SKDE is a very efficient algorithm giving compact representations and good approximations of the density functions. The second approach, based on Bayesian Networks, models the dynamics of the movement. Instead of considering motion patterns in each block independently, temporal sequences of motion patterns are learned by using Hidden Markov Models (HMM). The second proposed improvement consists in modeling the movement in one block by taking into account the observed motion in adjacent blocks. This is performed by the coupled HMM method. Evaluations were conducted to highlight the classification performance of the proposed methods,on both synthetic data and very challenging real video sequences captured by video surveillance cameras.These evaluations allow us to give first conclusions concerning automatic analyses of possibly crowded area.
|
252 |
[en] SEISMIC TO FACIES INVERSION USING CONVOLVED HIDDEN MARKOV MODEL / [pt] INVERSAO SÍSMICA PARA FÁCIES USANDO MODELO DE MARKOV OCULTO COM EFEITO CONVOLUTIVOERICK COSTA E SILVA TALARICO 07 January 2019 (has links)
[pt] A indústria de óleo e gás utiliza a sísmica para investigar a distribuição
de tipos de rocha (facies) em subsuperfície. Por outro lado, apesar de seu
corriqueiro uso em geociências, medidas sísmicas costumam ser ruidosas, e
a inversão do dado sísmico para a distribuição de facies é um problema mal
posto. Por esta razão, diversos autores estudam esta inversão sob o ponto
de vista probabilístico, para ao menos estimar as incertezas da solução do
problema inverso. O objetivo da presente dissertação é desenvolver método
quantitativo para estimar a probabilidade de reservatório com hidrocarboneto,
dado um traço sísmico de reflexão, integrando modelagem sísmica
direta, e conhecimento geológico a priori. Utiliza-se, um dos métodos mais
recentes para resolver o problema inverso: Modelo de Markov Oculto com
Efeito Convolucional (mais especificamente, a Aproximação por Projeção
de (1)). É demonstrado que o método pode ser reformulado em termos do
Modelo de Markov Oculto (MMO) ordinário. A teoria de sísmica de AVA
é apresentada, e usada conjuntamente com MMO com Efeito Convolucional
para resolver a inversão de sísmica para facies. A técnica de inversão
é avaliada usando-se medidas difundidas em Aprendizado de Máquina, em
um conjunto de experimentos variados e realistas. Apresenta-se uma técnica
para medir a capacidade do algoritmo em estimar valores confiáveis
de probabilidade. Pelos testes realizados a aproximação por projeção apresenta
distorções de probabilidade inferiores a 5 por cento, tornando-a uma técnica
útil para a indústria de óleo e gás. / [en] Oil and Gas Industry uses seismic data in order to unravel the distribution
of rock types (facies) in the subsurface. But, despite its widespread use,
seismic data is noisy and the inversion from seismic data to the underlying
rock distribution is an ill-posed problem. For this reason, many authors
have studied the topic in a probabilistic formulation, in order to provide
uncertainty estimations about the solution of the inversion problem. The
objective of the present thesis is to develop a quantitative method to estimate
the probability of hydrocarbon bearing reservoir, given a seismic
reflection profile, and, to integrate geological prior knowledge with geophysical
forward modelling. One of the newest methods for facies inversion is
used: Convolved Hidden Markov Model (more specifically the Projection
Approximation from (1)). It is demonstrated how Convolved HMM can be
reformulated as an ordinary Hidden Markov Model problem (which models
geological prior knowledge). Seismic AVA theory is introduced, and used
with Convolved HMM theory to solve the seismic to facies problem. The
performance of the inversion technique is measured with common machine
learning scores, in a broad set of realistic experiments. The technique capability
of estimating reliable probabilities is quantified, and it is shown
to present distortions smaller than 5 percent. As a conclusion, the studied Projection
Approximation is applicable for risk management in Oil and Gas
applications, which integrates geological and geophysical knowledge.
|
253 |
Recognition of Structures, Functions and Interactions of Proteins of Pathogens : Implications in Drug DiscoveryRamkrishnan, Gayatri January 2016 (has links) (PDF)
Significant advancements in genome sequencing techniques and other high-throughput initiatives have resulted in the availability of complete sequences of genomes of a large number of organisms, which provide an opportunity to study detailed biological information encoded therein. Identification of functional roles of proteins can aid in comprehension of various cellular activities in an organism, which is traditionally achieved using techniques pertaining to the field of molecular biology, protein chemistry and macromolecular crystallography. The established experimental methods for protein structure and function determination, although accurate and resourceful, are laborious and time consuming. Computational analyses of sequences of gene products and exploration of evolutionary relationships can give clues on protein structure and/or function with reasonable accuracy which can be used to direct experimental studies on proteins of interest, effectively. Moreover, with growing volumes of data, there has been a growing disparity in the number of well-characterized and uncharacterized proteins, further necessitating the use of computational methods for investigating evolutionary and structure-function relationships.
The remarkable progress made in the development of computational techniques (Chapter 1) has immensely contributed to the state-of-the-art biological sequence analysis and recognition of protein structure and function in a reliable manner. These methods have largely influenced the exploration of protein sequence-structure-function space.
One of the relevant applications of computational approaches is in the understanding of functional make-up of human pathogens, their complex interplay with the host and implications in pathogenesis. In this thesis, sensitive profile-based search procedures have been utilized to address various aspects in the context of three pathogens- Mycobacterium tuberculosis, Plasmodium falciparum and Trypanosoma brucei, which are causative agents of potentially life- threatening diseases. The existing drugs approved for the diseases, although of immense value in controlling the disease, have several shortcomings, the most important of them being the emergence of drug resistance that render the current treatment regimens futile. Thus, the identification of practicable targets and new drugs or new combination therapies become an important necessity. Analyses on structural and functional repertoire of proteins encoded in the pathogenic genomes can provide means for rational identification of therapeutic intervention strategies.
This thesis begins with the computational analyses of proteins encoded in M. tuberculosis genome. M. tuberculosis is a primary aetiological agent of tuberculosis in humans, and is o responsible for an estimated 1.5 million deaths every year. The complete genome of the pathogen was sequenced and made available more than a decade ago, which has been valuable in determination of functional roles of its gene products. Yet, functions of many M. tuberculosis proteins remain unknown. Computational prediction of protein function is an on- going process based on ever growing information made available in public databases as well as the introduction of powerful homology recognition techniques. Hence, a continuous refinement is essential to make the most of the sequence data, ensuring its accuracy and relevance. With the use of multiple sequence and structural profile-based search procedures, an enhanced structural and functional characterization of M. tuberculosis proteins, totalling to 95% of the genome was achieved (Chapter 2). Following are the key findings.
o Domain definitions were obtained for a total of 3566 of 4018 proteins. Amino acid residue coverage of >70% was achieved for 2295 proteins which constitute more than half of the proteome.
o Domain assignments were newly identified for 244 proteins with domain-unassigned regions. Structure prediction for these proteins corroborated all the remote homologyrelationships recognized using profile-based methods, enhancing the reliability of the predictions.
o Comparison on domain compositions of proteins between M. tuberculosis and human host, revealed presence of pathogen-specific domains that are not homologous to proteins in human. Such proteins in M. tuberculosis are mainly virulence factors involved in host-pathogen interactions such as immune-dominance and aiding entry and survival in human host macrophages, hence forming attractive targets for drug discovery.
o Putative structural and functional information for proteins with no recognizable domains were inferred by means of fold recognition and an iterative profile-based search against sequence database.
o Attributing putative structures and functions to 955 conserved hypothetical proteins in M. tuberculosis, 137 of which are reportedly essential to the pathogen, provide a basis to re-investigate their involvement in pathogenesis and survival in the host.
Proteins with no detectable homologues were recognized as M. tuberculosis H37Rv-specific, which can serve as promising drug targets.
An attempt was made to identify porin-like proteins in M. tuberculosis, considering MspA porin from M. smegmatis as a template. The difficulty in recognition of putative porins in M. tuberculosis is indicative of novel outer membrane channel proteins, not characterized yet, or high representation of ion-channels, symporters and transporters to compensate for the functional role of porins. In addition, MspA-like proteins were not readily recognized in other slow-growing mycobacterial pathogens that are known to infect human host, apart from M. tuberculosis. This indicates probable acquisition of physiological adaptations, i.e. absence of porins, to confer drug-resistance, in the course of their co-evolution with human hosts.
Evolutionary relationships recognized between sequence (Pfam) and structural (SCOP) families aided in association of potential structures and/or functions for 55 uncharacterized Pfam domains recognized in M. tuberculosis. Such associations deliver useful insights into the structure and function of a protein housing the uncharacterized domain.
The functional inferences drawn for M. tuberculosis proteins based on the predictions can provide valuable basis for experimental endeavours in understanding mechanisms of pathogenesis and can significantly impact anti-tubercular drug discovery programmes.
An interesting outcome benefitted from the exercise of exploring relationships between Pfam and SCOP families, was the identification of evolutionary relationship between a Pfam domain of unknown function DUF2652 and class III nucleotidyl cyclases. A detailed investigation was undertaken to assess this relationship (Chapter 3). Nucleotidyl cyclases synthesize cyclic nucleotides which are critical second messengers in signalling pathways. The DUF2652 family predominantly comprises of bacterial proteins belonging to three lineages- Actinobacteria, Bacteroidetes and Proteobacteria. Thus, recognition of evolutionary relationship between these bacterial proteins and nucleotide cyclases is of particular interest due to the indispensability of cyclic nucleotides in regulation of varied biological activities in bacteria. Use of fold recognition program suggested presence of nucleotide cyclase-characteristic topological motif (βααββαβ) in all the members of the DUF2652 family. Detailed analyses on structural and functional features of the uncharacterized set of bacterial proteins corresponding to 50 bacterial genomes, using profile- based alignments, revealed presence of key features typical of nucleotidyl cyclases, including metal-binding aspartates, substrate-specifying residues and transition-state stabilizing residues. Depending on the features, 20 proteins of Actinobacteria lineage, predominantly mycobacteria, of unknown structure and function were identified as putative nucleotide cyclases, 23 proteins of Bacteroidetes lineage were associated with guanylyl cyclases, while 8 uncharacterized proteins of Proteobacteria were recognized as nucleotide cyclase-like proteins (7 adenylyl and one guanylyl cyclase). Sequence similarity-based clustering of the predicted nucleotide cyclase-like proteins with established nucleotide cyclases indicated the apparent evolutionarily distinctness of the subfamily of class III nucleotidyl cyclases predicted. Furthermore, analysis of evolutionarily conserved gene clusters of the predicted nucleotide cyclase-like proteins indicated functional associations that support the predictions on their participation in cellular signalling events. The inferences made can be experimentally investigated further to ascertain the involvement of the uncharacterized bacterial proteins in signalling pathways, which can help in understanding the pathobiology of pathogenic species of interest.
The next objective was the recognition of biologically relevant protein-protein interactions across M. tuberculosis and human host (Chapter 4). M. tuberculosis is well known for its ability to successfully co-evolve with human host in terms of establishing infection, survival and persistence. The current knowledge on the mechanisms of host invasion, immune evasion and persistence in the host environment can be attributed, and is limited, to the experimental studies pursued by numerous groups. Chapter 4 presents an approach for computational identification of biologically feasible protein-protein interactions across M. tuberculosis and human host. The approach utilizes crystal structures of intra-organism protein-protein complexes which are transient in nature. Identification of homologues of host and pathogen proteins in the database of known protein-protein interactions, formed the initial step, followed by identification of conserved interfacial patch and integration of information on tissue-specific expression of human proteins and subcellular localization of human and M. tuberculosis proteins. In addition, appropriate filters were used to extract biologically feasible host-pathogen protein-protein interactions. This resulted in recognition of 386 interactions potentially mediated by 59 M. tuberculosis proteins and 90 human proteins. A predominance of host-pathogen interactions (193 protein-protein interactions) brought about by M. tuberculosis proteins participating in cell wall processes, was observed, which is in concurrence with the experimental studies on immuno-modulatory activities brought about by such proteins. These set of mycobacterial proteins were predicted to interact with diverse set of host proteins such as those involved in ubiquitin conjugation pathways, metabolic pathways, signalling pathways, regulation of cell proliferation, transport, apoptosis and autophagy. The predictions have the potential to complement experimental observations at the molecular level. Details on couple of interesting cases are presented in the chapter, one of which is the probable mechanism of immune evasion adopted by M. tuberculosis to inhibit lysozyme activity in macrophages, and second is the mechanism of nutrient uptake from host. The set of M. tuberculosis proteins predicted to mediate interactions with host proteins have the potential to warrant an experimental follow-up on probable mechanisms of pathogenesis and also serve as attractive targets for chemotherapeutic interventions.
proteins known to participate in P. falciparum metabolism. Pathway holes, where evidence on metabolic step exists but the catalysing enzyme is not known, have also been addressed in the study, several of which have been suggested to play an important role in growth and development of the parasite during its intra-erythrocytic stages in human host.
A subsequent objective was the recognition P. falciparum proteins potentially capable of remodelling erythrocytes to suit their niche (Chapter 7). Exploitative mechanisms are brought about by the parasite to remodel erythrocytes for growth and survival during intra-erythrocytic stages of its life-cycle, the understanding of which is limited to experimental studies. To achieve physicochemically viable protein-protein interactions potentially mediated by proteins of human erythrocytes and P. falciparum proteins, a structure-influenced protocol, similar to the one demonstrated in Chapter 4, was employed. Information on subcellular localization and protein expression is crucial especially for parasites like P. falciparum, which reside in One of the major shortcomings with current treatment regimen for tuberculosis is the emergence of multidrug (MDR) and extensively drug-resistant (XDR) strains that render first-line and second-line drug treatments futile. This entails a need to explore target space in M. tuberculosis as well as explore the potential of existing drugs for repurposing against tuberculosis. A drug repurposing strategy i.e. exploring within-target-family selectivity of small molecules, has been implemented (Chapter 5) to contribute towards time and cost-saving anti-tubercular drug development efforts. With the use of profile-based search procedures, evolutionary relationships between targets (other than proteins of M. tuberculosis) of FDA-approved drugs and M. tuberculosis proteins were investigated. A key filter to exclude drugs capable of acting on human proteins substantially reduced the chances of obtaining anti-targets. Thus, total of 130 FDA-approved drugs were recognized that can be repurposed against 78 M. tuberculosis proteins, belonging to the functional categories- intermediary metabolism and respiration, information pathways, cell wall and cell processes and lipid metabolism. The catalogue of structure and function of M. tuberculosis proteins and their involvement in host-pathogen protein-protein interactions compiled from chapters 2 and 4 served as a guiding tool to explore the functional importance of targets identified. Many of the potential targets identified have been experimentally shown to be essential for growth and survival of the pathogen earlier, thus gaining importance in terms of pharmaceutical relevance. Polypharmacological drugs or drugs capable of acting of multiple targets were also identified (92 drugs) in the study. These drugs have the potential to stand tolerance against development of drug resistance in the pathogen. Comparative sequence and structure-based analysis of M. tuberculosis proteins homologous to known targets yielded credible inferences on putative binding sites of FDA-approved drugs in potential targets. Instances where information on binding sites could not be readily inferred from known targets, potentially druggable sites have been predicted. Comparison with earlier experimental studies that report anti-tubercular potential of several approved drugs enhanced the credibility of 74 of 130 FDA-approved drugs that can be readily prioritized for clinical studies. An additional exercise was pursued to identify prospective anti-tubercular agents by means of structural comparison between ChEMBL compounds and 130 FDA-approved drugs. Only those compounds were retained that showed considerably high structural similarity with approved drugs. Such compounds with minor changes in terms of physicochemical properties provide a basis for exploration of compounds that may exhibit higher affinities to bind to M. tuberculosis targets. The set of approved drugs recognized as repurpose-able candidates against tuberculosis, in concert with the structurally similar compounds, can significantly impact anti-tubercular drug development and drug discovery.
The next part of the thesis focuses on Plasmodium falciparum, an obligate intracellular protozoan parasite responsible for malaria. The parasite genome features unusual characteristics including abundance of low complexity regions and pronounced sequence divergence that render protein structure and function recognition difficult. The parasite also manifests remarkable plasticity in its metabolic organization throughout its developmental stages in two hosts-human and mosquito; thus obtaining an exhaustive list of metabolic proteins in the parasite gains importance. Considering the utility of multiple sensitive profile-based search approaches in enhanced annotation of M. tuberculosis genome, a similar exercise was employed to recognize potential metabolic proteins in P. falciparum (Chapter 6). A total of 172 metabolic proteins were identified as participants of 78 metabolic pathways, over and above 609heterogeneous environmental conditions at different stages in their lifecycle. Inclusion of such data aided in extraction of 208 biologically relevant protein-protein interactions potentially mediated by 59 P. falciparum proteins and 30 erythrocyte proteins. Host-parasite protein-protein interactions were predicted pertaining to several major strategies spanning intra-erythrocytic stages in P. falciparum pathogenesis including- gaining entry into the host erythrocytes (category: RBC invasion, protease), redirecting parasitic proteins to erythrocyte membrane (category: protein traffic), modulating erythrocyte machinery (category: rosette formation, putative adhesin, chaperone, kinase), evading immunity (category: immune evasion) and eventually egress (category: merozoite egress) to infect other uninfected erythrocytes. Elaborate means to analyse and evaluate the functional viability of a predicted interaction in terms of geometrical packing at the interfacial region, electrostatic complementarity of the interacting surfaces and interaction energies is also demonstrated. The protein-protein interactions, thus predicted between human erythrocytes and P. falciparum, have the potential to provide a useful basis in understanding probable mechanisms of pathogenesis, and indeed in pinning down attractive targets for antimalarial drug discovery.
The emergence of drug resistance against all known antimalarial agents, currently in use, necessitates discovery and development of either new antimalarial agents or unexplored combination of drugs that may not only reduce mortality and morbidity of malaria, but also reduce the risk of resistance to antimalarial drugs. In an attempt to contribute towards the same, Chapter 8 explores the established concept of within-target-family selectivity of small molecules to recognize antimalarial potential of the approved drugs. Eighty six FDA-approved drugs, predominantly constituted by antibacterial agents, were identified as feasible candidates for repurposing against 90 P. falciparum proteins. Most of the potential parasite targets identified are known to participate in housekeeping machinery, protein biosynthesis, metabolic pathways and cell growth and differentiation, and thus are pharmaceutically relevant. During intra-erythrocytic growth of P. falciparum, the parasite resides within the erythrocyte, within a protective encasing, known as parasitophorous vacuole. Hence a drug, intended to target a parasite protein residing in an organelle, must be sufficiently hydrophilic or hydrophobic to be able to permeate cell membranes and reach its site of activity. On the basis of lipophilicity of the drugs, a physical property determined experimentally, 57 of 86 FDA-approved drugs were recognized as feasible candidates for use against P. falciparum during the course of blood-stages of infection, which can be prioritized for antimalarial drug development programmes.
The final section of the thesis focuses on the protozoan parasite Trypanosoma brucei, a causative agent of African sleeping sickness (Chapter 9). This disease is endemic to sub-Saharan regions of Africa. Despite the availability of completely sequenced genome of T. brucei, structure and function for about 50% of the proteins encoded in the genome remain unknown. Absence of prophylactic chemotherapy and vaccine, compounded with emergence of drug-resistance renders anti-trypanosomal drug discovery challenging. Thus, considering the utility of frameworks established in earlier chapters for recognition of protein structure, function and drug-targets, similar steps were undertaken to understand functional repertoire of the parasite and use drug repurposing methods to accelerate anti-trypanosomal drug discovery efforts. Structures and functions were reliably recognized for 70% of the gene products (5894) encoded in T. brucei genome, with the use of multiple profile-based search procedures, coupled with information on presence of transmembrane domains and signal peptide cleavage sites. Consequently, a total of 282 uncharacterized T. brucei proteins could be newly coined as potential metabolic proteins. Integration of information on stage-specific expression profiles with Trypanosoma-specific and T-.brucei-specific proteins identified in the study, aided in pinning down potential attractive targets. Additionally, exploration of evolutionary relationships between targets of FDA-approved drugs and T. brucei proteins, 68 FDA-approved drugs were predicted as repurpose-able candidates against 42 potential T. brucei targets which primarily include proteins involved in regulatory processes and metabolism. Several targets predicted are reportedly essential in assisting the parasite to switch between differentiation forms (bloodstream and procyclic) in the course of its lifecycle. These targets are of high therapeutic relevance, hence the corresponding drug-target associations provide a useful resource for experimental endeavours.
In summary, this thesis presents computational analyses on three pathogenic genomes in terms of enhancing the understanding of functional repertoire of the pathogens, addressing metabolic pathway holes, exploring probable mechanisms of pathogenesis brought about by potential host-pathogen protein-protein interactions, and identifying feasible FDA-approved drug candidates to repurpose against the pathogens. The studies are pursued primarily by taking advantage of powerful homology-detection techniques and the ever-growing biological information made available in public databases. Indeed, the inferences drawn for the three pathogenic genomes serve an excellent resource for an experimental follow-up. The set of protocols presented in the thesis are highly generic in nature, as demonstrated for three pathogens, and can be utilized for genome-wide analyses on many other pathogens of interest.
The supplemental data associated with the chapters is provided in a compact disc attached with this thesis.
|
254 |
Detecção de situações anormais em caldeiras de recuperação química. / Detection of abnormal situations in chemical recovery boilers.Gustavo Matheus de Almeida 12 September 2006 (has links)
O desafio para a área de monitoramento de processos, em indústrias químicas, ainda é a etapa de detecção, com a necessidade de desenvolvimento de sistemas confiáveis. Pode-se resumir que um sistema é confiável, ao ser capaz de detectar as situações anormais, de modo precoce, e, ao mesmo tempo, de minimizar a geração de alarmes falsos. Ao se ter um sistema confiável, pode-se empregá-lo para auxiliar o operador, de fábricas, no processo de tomada de decisões. O objetivo deste estudo é apresentar uma metodologia, baseada na técnica, modelo oculto de Markov (HMM, acrônimo de ?Hidden Markov Model?), para se detectar situações anormais em caldeiras de recuperação química. As aplicações de maior sucesso de HMM são na área de reconhecimento de fala. Pode-se citar como aspectos positivos: o raciocínio probabilístico, a modelagem explícita, e a identificação a partir de dados históricos. Fez-se duas aplicações. O primeiro estudo de caso é no ?benchmark? de um sistema de evaporação múltiplo efeito de uma fábrica de produção de açúcar. Identificou-se um HMM, característico de operação normal, para se detectar cinco situações anormais no atuador responsável por regular o fluxo de xarope de açúcar para o primeiro evaporador. A detecção, para as três situações abruptas, é imediata, uma vez que o HMM foi capaz de detectar alterações, abruptas, no sinal da variável monitorada. Em relação às duas situações incipientes, foi possível detectá-las ainda em estágio inicial; ao ser o valor de f (vetor responsável por representar a intensidade de um evento anormal, com o tempo), no instante da detecção, próximo a zero, igual a 2,8% e 2,1%, respectivamente. O segundo estudo de caso é em uma caldeira de recuperação química, de uma fábrica de produção de celulose, no Brasil. O objetivo é monitorar o acúmulo de depósitos de cinzas sobre os equipamentos da sessão de transferência de calor convectivo, através de medições de perda de carga. Este é um dos principais desafios para se aumentar a eficiência operacional deste equipamento. Após a identificação de um HMM característico de perda de carga alta, pôde-se verificar a sua capacidade de informar o estado atual e, por consequência, a tendência do sistema, de modo similar à um preditor. Pôde-se demonstrar também a utilidade de se definir limites de controle, com o objetivo de se ter a informação sobre a distância entre o estado atual e os níveis de alarme de perda de carga. / The greatest challenge faced by the area of process monitoring in chemical industries still resides in the fault detection task, which aims at developing reliable systems. One may say that a system is reliable if it is able to perform early fault detection and, at the same time, to reduce the generation of false alarms. Once there is a reliable system available, it can be employed to help operators, in factories, in the decisionmaking process. The aim of this study is presenting a methodology, based on the Hidden Markov Model (HMM) technique, suggesting its use in the detection of abnormal situations in chemical recovery boilers. The most successful applications of HMM are in the area of speech recognition. Some of its advantages are: probabilistic reasoning, explicit modeling and the identification based on process history data. This study discusses two applications. The first one is on a benchmark of a multiple evaporation system in a sugar factory. A HMM representative of the normal operation was identified, in order to detect five abnormal situations at the actuator responsible for controlling the syrup flow to the first evaporator. The detection result for the three abrupt situations was immediate, since the HMM was capable of detecting the statistical changes on the signal of the monitored variable as soon as they occurred. Regarding to the two incipient situations, the detection was done at an early stage. For both events, the value of vector f (responsible for representing the strength of an abnormal event over time), at the time it occurred, was near zero, equal to 2.8 and 2.1%, respectively. The second case study deals with the application of HMM in a chemical recovery boiler, belonging to a cellulose mill, in Brazil. The aim is monitoring the accumulation of ash deposits over the equipments of the convective heat transfer section, through pressure drop measures. This is one of the main challenges to be overcome nowadays, bearing in mind the interest that exists in increasing the operational efficiency of this equipment. Initially, a HMM for high values of pressure drop was identified. With this model, it was possible to check its capacity to inform the current state, and consequently, the tendency of the system (similarly as a predictor). It was also possible to show the utility of defining control limits, in order to inform the operator the relative distance between the current state of the system and the alarm levels of pressure drop.
|
255 |
Classificação de sinais EGG combinando Análise em Componentes Independentes, Redes Neurais e Modelo Oculto de MarkovSantos, Hallan Cosmo dos 26 May 2015 (has links)
Identify some digestive features in people through Electrogastrogram (EGG) is important because this is a cheap, non-invasive and less bother way than traditional endoscopy procedure. This work evaluates the learning behavior of Artificial Neural Networks (ANN) and Hidden Markov Model (HMM) on components extracted by Independent Component Analysis (ICA) algorithms. In this research, an experiment was made with statistical analysis that shows the relationship between neutral, negative or positive images and digestive reactions.
Training some classifiers with an EGG signal database, where the emotional states of individuals are known during processing, would it be possible to carry out the other way? Meaning, just from the EGG signal, estimate the emotional state of individuals. The initial challenge is to treat the EGG signal, which is mixed with the signals from other organs such as heart and lung. For this, the FastICA and Tensorial Methods algorithms were used, in order to produce a set of independent components, where one can identify the stomach component. Then, the EGG signal classification is performed through ANN and HMM models. The results have shown that extracting only the stomach signal component before the experiment can reduce the learning error rate in classifiers. / Identificar características digestivas de pessoas através da Eletrogastrografia (EGG) é importante pois esta costuma ser uma opção barata, não-invasiva e incomoda menos que o tradicional procedimento de Endoscopia. Este trabalho avalia o comportamento do aprendizado das Redes Neurais Artificiais (RNA) e do Modelo Oculto de Markov (HMM) diante de componentes extraídas por algoritmos de Análise de Componentes Independentes (ICA). Nesta pesquisa é realizado um experimento com análise estatística cujo objetivo apresenta a relação entre a visualização de imagens neutras, negativas ou positivas e as reações digestivas.
Treinando alguns classificadores com uma base de dados de sinais EGG, onde se conhece os estados emocionais dos indivíduos durante a sua obtenção, seria possível realizar o caminho inverso? Em outras palavras, apenas a partir dos sinais EGG, pode-se estimar o estado emocional de indivíduos? O desafio inicial é tratar o sinal EGG que encontra-se misturado aos sinais de outros órgãos como coração e pulmão. Para isto foi utilizado o algoritmo FastICA e os métodos tensoriais, com o intuito de produzir um conjunto de componentes independentes onde se possa identificar a componente do estômago. Em seguida, a classifição do sinal EGG é realizada por meio dos modelos de RNA e HMM. Os resultados mostraram que classificar apenas as componentes com mais presença da frequência do sinal do estômago pode reduzir a taxa de erro do aprendizado dos classificadores no experimento realizado.
|
256 |
Modélisation de signaux temporels hautes fréquences multicapteurs à valeurs manquantes : Application à la prédiction des efflorescences phytoplanctoniques dans les rivières et les écosystèmes marins côtiers / Modelling of high frequency time signals, multisensors with missing values : predicting application to algal blooms in rivers and coastal aquatic ecosystemsRousseeuw, Kévin 11 December 2014 (has links)
La prise de conscience des problèmes d'environnement et des effets directs et indirects des activités humaines a conduit à renforcer la surveillance haute fréquence des écosystèmes marins par l'installation de stations de mesures multicapteurs autonomes. Les capteurs, installés dans des milieux hostiles, sont sujets à des périodes de calibration, d'entretien voire des pannes et sont donc susceptibles de générer des données bruitées, manquantes voire aberrantes qu'il est nécessaire de filtrer et compléter avant toute exploitation ultérieure. Dans ce contexte, l'objectif du travail est de concevoir un système numérique automatisé robuste capable de traiter de tel volume de données afin d’améliorer les connaissances sur la qualité des systèmes aquatiques, et plus particulièrement en considérant le déterminisme et la dynamique des efflorescences du phytoplancton. L'étape cruciale est le développement méthodologique de modèles de prédiction des efflorescences du phytoplancton permettant aux utilisateurs de disposer de protocoles adéquats. Nous proposons pour cela l'emploi du modèle de Markov caché hybridé pour la détection et la prédiction des états de l'environnement (caractérisation des phases clefs de la dynamique et des caractéristiques hydrologiques associées). L'originalité du travail est l'hybridation du modèle de Markov par un algorithme de classification spectrale permettant un apprentissage non supervisé conjoint de la structure, sa caractérisation et la dynamique associée. Cette approche a été appliquée sur trois bases de données réelles : la première issue de la station marine instrumentée MAREL Carnot (Ifremer) (2005-2009), la seconde d’un système de type Ferry Box mis en œuvre en Manche orientale en 2012 et la troisième d’une station de mesures fixe, installée le long de la rivière Deûle en 2009 (Agence de l’Eau Artois Picardie - AEAP). Le travail s’inscrit dans le cadre d’une collaboration étroite entre l'IFREMER, le LISIC/ULCO et l'AEAP afin de développer des systèmes optimisés pour l’étude de l’effet des activités anthropiques sur le fonctionnement des écosystèmes aquatiques et plus particulièrement dans le contexte des efflorescences de l’algue nuisible, Phaeocystis globosa. / Because of the growing interest for environmental issues and to identify direct and indirect effects of anthropogenic activities on ecosystems, environmental monitoring programs have recourse more and more frequently to high resolution, autonomous and multi-sensor instrumented stations. These systems are implemented in harsh environment and there is a need to stop measurements for calibration, service purposes or just because of sensors failure. Consequently, data could be noisy, missing or out of range and required some pre-processing or filtering steps to complete and validate raw data before any further investigations. In this context, the objective of this work is to design an automatic numeric system able to manage such amount of data in order to further knowledge on water quality and more precisely with consideration about phytoplankton determinism and dynamics. Main phase is the methodological development of phytoplankton bloom forecasting models giving the opportunity to end-user to handle well-adapted protocols. We propose to use hybrid Hidden Markov Model to detect and forecast environment states (identification of the main phytoplankton bloom steps and associated hydrological conditions). The added-value of our approach is to hybrid our model with a spectral clustering algorithm. Thus all HMM parameters (states, characterisation and dynamics of these states) are built by unsupervised learning. This approach was applied on three data bases: first one from the marine instrumented station MAREL Carnot (Ifremer) (2005-2009), second one from a Ferry Box system implemented in the eastern English Channel en 2012 and third one from a freshwater fixed station in the river Deûle in 2009 (Artois Picardie Water Agency). These works fall within the scope of a collaboration between IFREMER, LISIC/ULCO and Artois Picardie Water Agency in order to develop optimised systems to study effects of anthropogenic activities on aquatic systems functioning in a regional context of massive blooms of the harmful algae, Phaeocystis globosa.
|
257 |
Étude des fonctions B-splines pour la fusion d'images segmentées par approche bayésienne / Study of B-spline function for fusion of segmented images by Bayesian approachHadrich Ben Arab, Atizez 02 December 2015 (has links)
Dans cette thèse nous avons traité le problème de l'estimation non paramétrique des lois de probabilités. Dans un premier temps, nous avons supposé que la densité inconnue f a été approchée par un mélange de base B-spline quadratique. Puis, nous avons proposé un nouvel estimateur de la densité inconnue f basé sur les fonctions B-splines quadratiques, avec deux méthodes d'estimation. La première est base sur la méthode du maximum de vraisemblance et la deuxième est basée sur la méthode d'estimation Bayésienne MAP. Ensuite, nous avons généralisé notre étude d'estimation dans le cadre du mélange et nous avons proposé un nouvel estimateur du mélange de lois inconnues basé sur les deux méthodes d'estimation adaptées. Dans un deuxième temps, nous avons traité le problème de la segmentation statistique semi supervisée des images en se basant sur le modèle de Markov caché et les fonctions B-splines. Nous avons montré l'apport de l'hybridation du modèle de Markov caché et les fonctions B-splines en segmentation statistique bayésienne semi supervisée des images. Dans un troisième temps, nous avons présenté une approche de fusion basée sur la méthode de maximum de vraisemblance, à travers l'estimation non paramétrique des probabilités, pour chaque pixel de l'image. Nous avons ensuite appliqué cette approche sur des images multi-spectrales et multi-temporelles segmentées par notre algorithme non paramétrique et non supervisé. / In this thesis we are treated the problem of nonparametric estimation probability distributions. At first, we assumed that the unknown density f was approximated by a basic mixture quadratic B-spline. Then, we proposed a new estimate of the unknown density function f based on quadratic B-splines, with two methods estimation. The first is based on the maximum likelihood method and the second is based on the Bayesian MAP estimation method. Then we have generalized our estimation study as part of the mixture and we have proposed a new estimator mixture of unknown distributions based on the adapted estimation of two methods. In a second time, we treated the problem of semi supervised statistical segmentation of images based on the hidden Markov model and the B-sline functions. We have shown the contribution of hybridization of the hidden Markov model and B-spline functions in unsupervised Bayesian statistical image segmentation. Thirdly, we presented a fusion approach based on the maximum likelihood method, through the nonparametric estimation of probabilities, for each pixel of the image. We then applied this approach to multi-spectral and multi-temporal images segmented by our nonparametric and unsupervised algorithm.
|
258 |
Estimation aveugle de chaînes de Markov cachées simples et doubles : Application au décodage de codes graphiques / Blind estimation of hidden and double Markov chain : Application to barcode decodingDridi, Noura 25 June 2012 (has links)
Depuis leur création, les codes graphiques constituent un outil d'identification automatique largement exploité en industrie. Cependant, les performances de lecture sont limitées par un flou optique et un flou de mouvement. L'objectif de la thèse est l'optimisation de lecture des codes 1D et 2D en exploitant des modèles de Markov cachés simples et doubles, et des méthodes d'estimation aveugles. En premier lieu, le système de lecture de codes graphiques est modélisé par une chaîne de Markov cachée, et des nouveaux algorithmes pour l'estimation du canal et la détection des symboles sont développés. Ils tiennent compte de la non stationnarité de la chaîne de Markov. De plus une méthode d'estimation de la taille du flou et de sa forme est proposée. La méthode utilise des critères de sélection permettant de choisir le modèle de dégradation le plus adéquat. Enfin nous traitons le problème de complexité qui est particulièrement important dans le cas d'un canal à mémoire longue. La solution proposée consiste à modéliser le canal à mémoire longue par une chaîne de Markov double. Sur la base de ce modèle, des algorithmes offrant un rapport optimisé performance-complexité sont présentés / Since its birth, the technology of barcode is well investigated for automatic identification. When reading, a barcode can be degraded by a blur , caused by a bad focalisation and/ or a camera movement. The goal of this thesis is the optimisation of the receiver of 1D and 2D barcode from hidden and double Markov model and blind statistical estimation approaches. The first phase of our work consists of modelling the original image and the observed one using Hidden Markov model. Then, new algorithms for joint blur estimation and symbol detection are proposed, which take into account the non-stationarity of the hidden Markov process. Moreover, a method to select the most relevant model of the blur is proposed, based on model selection criterion. The method is also used to estimate the blur length. Finally, a new algorithm based on the double Markov chain is proposed to deal with digital communication through a long memory channel. Estimation of such channel is not possible using the classical detection algorithms based on the maximum likelihood due to the prohibitive complexity. New algorithm giving good trade off between complexity and performance is provided
|
259 |
Méthodes de Monte-Carlo EM et approximations particulaires : application à la calibration d'un modèle de volatilité stochastique / Monte Carlo EM methods and particle approximations : application to the calibration of stochastic volatility modelAllaya, Mouhamad M. 09 December 2013 (has links)
Ce travail de thèse poursuit une perspective double dans l'usage conjoint des méthodes de Monte Carlo séquentielles (MMS) et de l'algorithme Espérance-Maximisation (EM) dans le cadre des modèles de Markov cachés présentant une structure de dépendance markovienne d'ordre supérieur à 1 au niveau de la composante inobservée. Tout d'abord, nous commençons par un exposé succinct de l'assise théorique des deux concepts statistiques à Travers les chapitres 1 et 2 qui leurs sont consacrés. Dans un second temps, nous nous intéressons à la mise en pratique simultanée des deux concepts au chapitre 3 et ce dans le cadre usuel ou la structure de dépendance est d'ordre 1, l'apport des méthodes MMS dans ce travail réside dans leur capacité à approximer efficacement des fonctionnelles conditionnelles bornées, notamment des quantités de filtrage et de lissage dans un cadre non linéaire et non gaussien. Quant à l'algorithme EM, il est motivé par la présence à la fois de variables observables, et inobservables (ou partiellement observées) dans les modèles de Markov Cachés et singulièrement les modèles de volatilité stochastique étudié. Après avoir présenté aussi bien l'algorithme EM que les méthodes MCS ainsi que quelques une de leurs propriétés dans les chapitres 1 et 2 respectivement, nous illustrons ces deux outils statistiques au travers de la calibration d'un modèle de volatilité stochastique. Cette application est effectuée pour des taux change ainsi que pour quelques indices boursiers au chapitre 3. Nous concluons ce chapitre sur un léger écart du modèle de volatilité stochastique canonique utilisé ainsi que des simulations de Monte Carlo portant sur le modèle résultant. Enfin, nous nous efforçons dans les chapitres 4 et 5 à fournir les assises théoriques et pratiques de l'extension des méthodes Monte Carlo séquentielles notamment le filtrage et le lissage particulaire lorsque la structure markovienne est plus prononcée. En guise d’illustration, nous donnons l'exemple d'un modèle de volatilité stochastique dégénéré dont une approximation présente une telle propriété de dépendance. / This thesis pursues a double perspective in the joint use of sequential Monte Carlo methods (SMC) and the Expectation-Maximization algorithm (EM) under hidden Markov models having a Markov dependence structure of order grater than one in the unobserved component signal. Firstly, we begin with a brief description of the theoretical basis of both statistical concepts through Chapters 1 and 2 that are devoted. In a second hand, we focus on the simultaneous implementation of both concepts in Chapter 3 in the usual setting where the dependence structure is of order 1. The contribution of SMC methods in this work lies in their ability to effectively approximate any bounded conditional functional in particular, those of filtering and smoothing quantities in a non-linear and non-Gaussian settings. The EM algorithm is itself motivated by the presence of both observable and unobservable ( or partially observed) variables in Hidden Markov Models and particularly the stochastic volatility models in study. Having presented the EM algorithm as well as the SMC methods and some of their properties in Chapters 1 and 2 respectively, we illustrate these two statistical tools through the calibration of a stochastic volatility model. This application is clone for exchange rates and for some stock indexes in Chapter 3. We conclude this chapter on a slight departure from canonical stochastic volatility model as well Monte Carlo simulations on the resulting model. Finally, we strive in Chapters 4 and 5 to provide the theoretical and practical foundation of sequential Monte Carlo methods extension including particle filtering and smoothing when the Markov structure is more pronounced. As an illustration, we give the example of a degenerate stochastic volatility model whose approximation has such a dependence property.
|
260 |
Modèles de Mobilité de Véhicules par Apprentissage Profond dans les Systèmes de Tranport Intelligents / Deep Learning based Vehicular Mobility Models for Intelligent Transportation SystemsZhang, Jian 07 December 2018 (has links)
Les systèmes de transport intelligents ont acquis un grand intérêt pour la recherche ces dernières années. Alors que la simulation réaliste du trafic joue un rôle important, elle n'a pas reçu suffisamment d'attention. Cette thèse est consacrée à l'étude de la simulation du trafic au niveau microscopique et propose des modèles de mobilité des véhicules correspondants. À l'aide de méthodes d'apprentissage profond, ces modèles de mobilité ont fait leurs preuves avec une crédibilité prometteuse pour représenter les véhicules dans le monde réel. D'abord, un modèle de mobilité basé sur un réseau de neurones piloté par les données est proposé. Ce modèle provient de données de trajectoires du monde réel et permet de mimer des comportements de véhicules locaux. En analysant les performances de ce modèle de mobilité basé sur un apprentissage de base, nous indiquons qu’une amélioration est possible et proposons ses spécifications. Un MMC est alors introduit. La préparation de cette intégration est nécessaire, ce qui comprend un examen des modèles de mobilité traditionnels basés sur la dynamique et l’adaptation des modèles « classiques » à notre situation. Enfin, le modèle amélioré est présenté et une simulation de scénarios sophistiqués est construite pour valider les résultats théoriques. La performance de notre modèle de mobilité est prometteuse et des problèmes de mise en œuvre sont également discutés / The intelligent transportation systems gain great research interests in recent years. Although the realistic traffic simulation plays an important role, it has not received enough attention. This thesis is devoted to studying the traffic simulation in microscopic level, and proposes corresponding vehicular mobility models. Using deep learning methods, these mobility models have been proven with a promising credibility to represent the vehicles in real-world. Firstly, a data-driven neural network based mobility model is proposed. This model comes from real-world trajectory data and allows mimicking local vehicle behaviors. By analyzing the performance of this basic learning based mobility model, we indicate that an improvement is possible and we propose its specification. An HMM is then introduced. The preparation of this integration is necessary, which includes an examination of traditional dynamics based mobility models and the adaptation method of “classical” models to our situation. At last, the enhanced model is presented, and a sophisticated scenario simulation is built with it to validate the theoretical results. The performance of our mobility model is promising and implementation issues have also been discussed
|
Page generated in 0.0603 seconds