Дисертації з теми "PLSDA"

La recherche de nouveaux composés pour prévenir ou atténuer le vieillissement de la peau est une priorité des recherches actuelles dans les cosmétiques. Dans ce contexte, le venin de frelon asiatique (Vespa velutina nigrithorax) a été étudié comme une source particulière de molécules potentiellement bioactives d’intérêt dermacosmétique. La première étude a tout d’abord porté sur la mise en œuvre d’un protocole fiable d’extraction et récupération du venin. Puis, la fraction peptidique et petites molécules a été sélectionnée afin d’évaluer, en comparaison avec le venin brut, la présence de molécules actives vis-à-vis d’une activité antioxydante, anti-microbienne (C. acnes) et inhibitrice enzymatique (tyrosinase, élastase, collagénase) in-tubo et in-cellulo. Ces études ont conduit à identifier par UHPLC-ESI-QTOF-HRMS/MS, dans le venin brut, une molécule responsable de l’activité anti-oxydante sur kératinocytes HaCaT. Dans une seconde étude, une approche peptidomique basée sur une méthode UHPLC-QTOF-HRMS et MS/MS suivie par un traitement statistique (PCA, PLS-DA) a été appliquée sur l’étude différentielle de profil peptidique du venin, en fonction de la période de collecte, des castes et du comportement. Ces derniers ont pour but d’évaluer l’influence de différents facteurs sur le patrimoine moléculaire de ces venins. Parallèlement, en troisième étude, une approche de criblage d’interaction Ligand/enzyme par spectrométrie de masse sur les enzymes élastase et tyrosinase immobilisées a été développée. Cette méthode a pour objectif de mettre en évidence la présence d’inhibiteurs ou de substrats dans des fractions plus ou moins complexes. On a montré que deux peptides présents dans le venin de frelon étaient capables d’interagir avec l’enzyme élastase en tant que substrat. La séquence peptidique de ces peptides a été partiellement obtenue par séquençage de novo
The search for new compounds to prevent or attenuate skin aging is a priority in current research in cosmetics. In this context, Asian Hornet venom (Vespa velutina nigrithorax) has been studied as a particular source of potentially bioactive molecules for dermacosmetic interest.The first study focused on the implementation of a reliable venom extraction and sampling protocol. Then, the peptide - small molecules fraction was selected to evaluate, in comparison with crude venom, the presence of active molecules with respect to antioxidant, anti-microbial (C. acnes) and enzyme inhibition (tyrosinase, elastase, collagenase) activity in-tubo and in-cellulo. These studies led to the identification in crude venom, by UHPLC-ESI-QTOF-HRMS/MS, of one molecule responsible for antioxidant activity on HaCaT keratinocytes.In a second study, a peptidomic approach based on UHPLC-ESI-QTOF-HRMS/MS followed by statistical processing (PCA, PLS-DA) was applied to the differential study of venom, according to the collection period, castes and behavior. The latter aims at evaluating the influence of these different factors on the venom molecular heritage. At the same time, in a third study, a ligand/enzyme interaction screening approach by mass spectrometry on solid-supported elastase enzymes was developed. The aim of this method is to detect the presence of inhibitors or substrates in more or less complex fractions. Two hornet venom peptides presenting in the hornet venom were identified to be capable of interacting with the enzyme elastase. Their peptide sequences were then partially obtained by de novo sequencing

Štavelj (Rumex crispus, Polygonaceae) je višegodišnja zeljasta biljka, koja predstavlja bogat izvor fenolnih komponenti. Iako se smatra invazivnim korovom, mlado lišće štavelja je jestivo i često se koristi kao salata. Dalje, upotreba plodova štavelja opisana je u srpskoj i turskoj narodnoj medicini u lečenju gastrointestinalnih tegoba. Cilj ovog rada bio je procena in vitro i in vivo antioksidantne/prooksidantne i citotoksične aktivnosti, i određivanje eventualnog in vitro antiinflamatornog efekta vodenog ekstrakta ploda Rumex crispus. Ukupan sadržaj flavonoida određen je spektrofotometrijskom metodom. Kvalifikacija i kvantifikacija flavonoida potvrđena je visokoefikasnom tečnom hromatografijom (HPLC). Antioksidantna aktivnost vodenog ekstrakta ploda štavelja procenjena je na osnovu in vitro testova: Ferric-reducing antioxidant power (FRAP), sposobnosti ekstrakta da neutrališe slobodne radikale NO•, OH• i DPPH• i uticaja na lipidnu peroksidaciju u lipozomima. Citotoksičnost ispitivanog ekstrakta je određena in vitro na tumorskim ćelijskim linijama: humani karcinom cerviksa (HeLa), adenokarcinom (HT-29) i adenokarcinom dojke (MCF7). Takođe, moguća in vivo hepatoprotektivna i antioksidantna svojstva ekstrakta određena su kod oksidativnog stresa izazvanog CCl4 kod eksperimentalnih životinja. Pored toga, proverena je hipoteza u kojoj testiran ekstrakt pokazuje in vivo antiproliferativnu aktivnost kod Ehrlich-ovih (EAC) i Hepatoma AS30D ćelija, merenjem zapremine ascitesa, procenta vijabilnih ćelija i nivoa nekoliko antioksidantnih enzima. Optimizovan in vitro test za određivanje potencijala inhibicije ciklooksigenaze-1 (COX-1) i 12-lipooksigenaze (12-LOX) preduzet je u svrhu procene antiinflamatornog efekta vodenog ekstrakta ploda R. crispus. HPLC analiza otkrila je da je mikvelianin najdominantniji flavonoidni konstituent ekstrakta. Testirani ekstrakt pokazao je potencijalnu antioksidantnu aktivnost rezultujući velikom moći u neutralizaciji slobodnih radikala, i sposobnošću da smanji lipidnu peroksidaciju u lipozomima. Rezultati su ukazali na tkivno-selektivnu citotoksičnost ekstrakta ploda R. crispus in vitro. Najizraženija antitumorska aktivnost primećena je prema HeLa i MCF7 ćelijskim linijama. Podaci sugerišu da bi se ispitivani ekstrakt mogao smatrati potencijalnim in vivo hepatoprotektivnim i antioksidantnim agensom, sprečavajući oksidativna oštećenja jetre. S druge strane, pomenuti ekstrakt može pokazati in vivo prooksidantna svojstva, uzrokujući oksidativni stres u maligno transformisanim EAC i AS30D ćelijama i smanjujući zapreminu ascitesa i udeo vijabilnih ćelija, u poređenju sa kontrolnom grupom. Promene u aktivnosti antioksidantnih enzima su verovatno posledica indukovanog oksidativnog stresa u EAC i AS30D ćelijama, naročito kod pretretiranih životinja. Vodeni ekstrakt ploda štavelja pokazao je COX-1, kao i 12-LOX inhibitornu aktivnost, navodeći da bi ispitivani ekstrakt mogao biti antiinflamatorni agens. Vodeni ekstrakt ploda R. crispus ima potencijalnu antioksidantnu, citotoksičnu i antiinflamatornu aktivnost. Ispoljavanje prooksidantnih svojstava predstavlja mogući mehanizam antiproliferativnog efekta ekstrakta.
Curly dock (Rumex crispus, Polygonaceae) is a wild perennial herbaceous plant, which products are described as a rich source of phenolic compounds. Apart from being considered a seriously invasive weed, young leaves of curly dock are edible and often used as salad. Furthermore, the use of its fruits has been described in Serbian and Turkish traditional medicine against stomach complaints. The objectives of this study were to evaluate in vitro and in vivo antioxidant/prooxidant and cytotoxic activities, and to determine an eventual in vitro anti-inflammatory effect of the aqueous extract of Rumex crispus fruits. Total flavonoid content was determined by spectrophotometric method. Qualification and quantification of flavonoids were confirmed using High performance liquid chromatography (HPLC). The aqueous extract of curly dock fruits was evaluated for its antioxidant activity by in vitro assays for Ferric-reducing antioxidant power (FRAP), NO•, OH• and DPPH•-free radical scavenging activities and the influence on lipid peroxidation in liposomes. The cytotoxicity of tested extract was examined in vitro in human cervix carcinoma (HeLa), colon adenocarcinoma (HT-29) and breast adenocarcinoma (MCF7). Also, the potential in vivo hepatoprotective and antioxidant properties of investigated extract were determined on CCl4-induced oxidative stress in experimental animals. Furthermore, the hypothesis that the examined extract might show in vivo antiproliferative activity in Ehrlich carcinoma (EAC) and Hepatoma AS30D cells was tested by measuring volume of ascites, percentage of viable cells and level of several antioxidant enzymes. The optimized in vitro test for determination of cyclooxygenase-1 (COX-1) and 12-lipoxygenase (12-LOX) inhibition potency was undertaken in order to estimate an anti-inflammatory effect of aqueous extract of R. crispus fruits. HPLC analysis revealed miquelianin as the most abundant flavonoid constituent of the extract. The tested extract might have an antioxidant activity resulting in scavenging of free radicals and ability to decrease lipid peroxidation in liposomes. The results could indicate tissue-selective cytotoxicity of R. crispus fruit extract in vitro. The most prominent antitumor activity was observed towards HeLa and MCF7 cell lines. The data suggested that investigated extract may be considered as potential in vivo hepatoprotective and antioxidant agent due to prevention of the liver injuries induced by oxidative damage. On the other hand, mentioned extract could exhibit in vivo prooxidant property, causing the oxidative stress in malignant transformed EAC and AS30D cells and reducing volume of ascites and percentage of viable cells, in comparison with control group. Changes in activities of antioxidant enzymes might be the results of induced oxidative stress in EAC and AS30D cells, especially in the pretreated animals. The aqueous extract of curly dock fruits showed COX-1, as well as 12-LOX inhibitory activity, suggesting that tested extract might be an anti-inflammatory agent. It could be concluded that aqueous fruit extract of R. crispus might have antioxidant, cytotoxic and anti-inflammatory activities. The prooxidant properties of examined extract could be the mechanism of potential antiproliferative effect of extract.

The present Ph.D. thesis, primarily conceived to support and reinforce the relation between academic and industrial worlds, was developed in collaboration with Shell Global Solutions (Amsterdam, The Netherlands) in the endeavour of applying and possibly extending well-established latent variable-based approaches (i.e. Principal Component Analysis - PCA - Partial Least Squares regression - PLS - or Partial Least Squares Discriminant Analysis - PLSDA) for complex problem solving not only in the fields of manufacturing troubleshooting and optimisation, but also in the wider environment of multivariate data analysis. To this end, novel efficient algorithmic solutions are proposed throughout all chapters to address very disparate tasks, from calibration transfer in spectroscopy to real-time modelling of streaming flows of data. The manuscript is divided into the following six parts, focused on various topics of interest: Part I - Preface, where an overview of this research work, its main aims and justification is given together with a brief introduction on PCA, PLS and PLSDA; Part II - On kernel-based extensions of PCA, PLS and PLSDA, where the potential of kernel techniques, possibly coupled to specific variants of the recently rediscovered pseudo-sample projection, formulated by the English statistician John C. Gower, is explored and their performance compared to that of more classical methodologies in four different applications scenarios: segmentation of Red-Green-Blue (RGB) images, discrimination of on-/off-specification batch runs, monitoring of batch processes and analysis of mixture designs of experiments; Part III - On the selection of the number of factors in PCA by permutation testing, where an extensive guideline on how to accomplish the selection of PCA components by permutation testing is provided through the comprehensive illustration of an original algorithmic procedure implemented for such a purpose; Part IV - On modelling common and distinctive sources of variability in multi-set data analysis, where several practical aspects of two-block common and distinctive component analysis (carried out by methods like Simultaneous Component Analysis - SCA - DIStinctive and COmmon Simultaneous Component Analysis - DISCO-SCA - Adapted Generalised Singular Value Decomposition - Adapted GSVD - ECO-POWER, Canonical Correlation Analysis - CCA - and 2-block Orthogonal Projections to Latent Structures - O2PLS) are discussed, a new computational strategy for determining the number of common factors underlying two data matrices sharing the same row- or column-dimension is described, and two innovative approaches for calibration transfer between near-infrared spectrometers are presented; Part V - On the on-the-fly processing and modelling of continuous high-dimensional data streams, where a novel software system for rational handling of multi-channel measurements recorded in real time, the On-The-Fly Processing (OTFP) tool, is designed; Part VI - Epilogue, where final conclusions are drawn, future perspectives are delineated, and annexes are included.
La presente tesis doctoral, concebida principalmente para apoyar y reforzar la relación entre la academia y la industria, se desarrolló en colaboración con Shell Global Solutions (Amsterdam, Países Bajos) en el esfuerzo de aplicar y posiblemente extender los enfoques ya consolidados basados en variables latentes (es decir, Análisis de Componentes Principales - PCA - Regresión en Mínimos Cuadrados Parciales - PLS - o PLS discriminante - PLSDA) para la resolución de problemas complejos no sólo en los campos de mejora y optimización de procesos, sino también en el entorno más amplio del análisis de datos multivariados. Con este fin, en todos los capítulos proponemos nuevas soluciones algorítmicas eficientes para abordar tareas dispares, desde la transferencia de calibración en espectroscopia hasta el modelado en tiempo real de flujos de datos. El manuscrito se divide en las seis partes siguientes, centradas en diversos temas de interés: Parte I - Prefacio, donde presentamos un resumen de este trabajo de investigación, damos sus principales objetivos y justificaciones junto con una breve introducción sobre PCA, PLS y PLSDA; Parte II - Sobre las extensiones basadas en kernels de PCA, PLS y PLSDA, donde presentamos el potencial de las técnicas de kernel, eventualmente acopladas a variantes específicas de la recién redescubierta proyección de pseudo-muestras, formulada por el estadista inglés John C. Gower, y comparamos su rendimiento respecto a metodologías más clásicas en cuatro aplicaciones a escenarios diferentes: segmentación de imágenes Rojo-Verde-Azul (RGB), discriminación y monitorización de procesos por lotes y análisis de diseños de experimentos de mezclas; Parte III - Sobre la selección del número de factores en el PCA por pruebas de permutación, donde aportamos una guía extensa sobre cómo conseguir la selección de componentes de PCA mediante pruebas de permutación y una ilustración completa de un procedimiento algorítmico original implementado para tal fin; Parte IV - Sobre la modelización de fuentes de variabilidad común y distintiva en el análisis de datos multi-conjunto, donde discutimos varios aspectos prácticos del análisis de componentes comunes y distintivos de dos bloques de datos (realizado por métodos como el Análisis Simultáneo de Componentes - SCA - Análisis Simultáneo de Componentes Distintivos y Comunes - DISCO-SCA - Descomposición Adaptada Generalizada de Valores Singulares - Adapted GSVD - ECO-POWER, Análisis de Correlaciones Canónicas - CCA - y Proyecciones Ortogonales de 2 conjuntos a Estructuras Latentes - O2PLS). Presentamos a su vez una nueva estrategia computacional para determinar el número de factores comunes subyacentes a dos matrices de datos que comparten la misma dimensión de fila o columna y dos planteamientos novedosos para la transferencia de calibración entre espectrómetros de infrarrojo cercano; Parte V - Sobre el procesamiento y la modelización en tiempo real de flujos de datos de alta dimensión, donde diseñamos la herramienta de Procesamiento en Tiempo Real (OTFP), un nuevo sistema de manejo racional de mediciones multi-canal registradas en tiempo real; Parte VI - Epílogo, donde presentamos las conclusiones finales, delimitamos las perspectivas futuras, e incluimos los anexos.
La present tesi doctoral, concebuda principalment per a recolzar i reforçar la relació entre l'acadèmia i la indústria, es va desenvolupar en col·laboració amb Shell Global Solutions (Amsterdam, Països Baixos) amb l'esforç d'aplicar i possiblement estendre els enfocaments ja consolidats basats en variables latents (és a dir, Anàlisi de Components Principals - PCA - Regressió en Mínims Quadrats Parcials - PLS - o PLS discriminant - PLSDA) per a la resolució de problemes complexos no solament en els camps de la millora i optimització de processos, sinó també en l'entorn més ampli de l'anàlisi de dades multivariades. A aquest efecte, en tots els capítols proposem noves solucions algorítmiques eficients per a abordar tasques dispars, des de la transferència de calibratge en espectroscopia fins al modelatge en temps real de fluxos de dades. El manuscrit es divideix en les sis parts següents, centrades en diversos temes d'interès: Part I - Prefaci, on presentem un resum d'aquest treball de recerca, es donen els seus principals objectius i justificacions juntament amb una breu introducció sobre PCA, PLS i PLSDA; Part II - Sobre les extensions basades en kernels de PCA, PLS i PLSDA, on presentem el potencial de les tècniques de kernel, eventualment acoblades a variants específiques de la recentment redescoberta projecció de pseudo-mostres, formulada per l'estadista anglés John C. Gower, i comparem el seu rendiment respecte a metodologies més clàssiques en quatre aplicacions a escenaris diferents: segmentació d'imatges Roig-Verd-Blau (RGB), discriminació i monitorització de processos per lots i anàlisi de dissenys d'experiments de mescles; Part III - Sobre la selecció del nombre de factors en el PCA per proves de permutació, on aportem una guia extensa sobre com aconseguir la selecció de components de PCA a través de proves de permutació i una il·lustració completa d'un procediment algorítmic original implementat per a la finalitat esmentada; Part IV - Sobre la modelització de fonts de variabilitat comuna i distintiva en l'anàlisi de dades multi-conjunt, on discutim diversos aspectes pràctics de l'anàlisis de components comuns i distintius de dos blocs de dades (realitzat per mètodes com l'Anàlisi Simultània de Components - SCA - Anàlisi Simultània de Components Distintius i Comuns - DISCO-SCA - Descomposició Adaptada Generalitzada en Valors Singulars - Adapted GSVD - ECO-POWER, Anàlisi de Correlacions Canòniques - CCA - i Projeccions Ortogonals de 2 blocs a Estructures Latents - O2PLS). Presentem al mateix temps una nova estratègia computacional per a determinar el nombre de factors comuns subjacents a dues matrius de dades que comparteixen la mateixa dimensió de fila o columna, i dos plantejaments nous per a la transferència de calibratge entre espectròmetres d'infraroig proper; Part V - Sobre el processament i la modelització en temps real de fluxos de dades d'alta dimensió, on dissenyem l'eina de Processament en Temps Real (OTFP), un nou sistema de tractament racional de mesures multi-canal registrades en temps real; Part VI - Epíleg, on presentem les conclusions finals, delimitem les perspectives futures, i incloem annexos.
Vitale, R. (2017). Novel chemometric proposals for advanced multivariate data analysis, processing and interpretation [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/90442
TESIS

Chemometrics is a chemical discipline that uses mathematical and statistical methods in order to extract useful information from multivariate chemical data. Moreover, chemometrics is applied to correlate quality parameters or physical properties to analytical instrument data such as calculating pH from a measurement of hydrogen ion activity or a Fourier transform interpolation of a spectrum. Aim of this thesis project is to develop chemometrical strategies for the elaboration and the interpretation of non-selective complex data in order to solve real problems in food, industry and environmental fields.

U okviru ove disertacije ispitan je hemijski, mineralni ipolifenolni sastav, kao i biološki potencijal četiri sortekupina sa dva različita lokaliteta sjeverozapadnog dijelaBosne i Hercegovine, sa lokaliteta Verići (divlja ikultivisana sorta Čačanska bestrna) i sa lokalitetaJavorani (divlja sorta i kultivisana sorta ChesterThornles). Određen je sadržaj suve materije, pepela,sirove celuloze, ukupnih šećera, ukupna kiselost i sadržajaskorbinske kiseline. Sadržaj mineralnih materijadetektovan je metodom optičke emisione spektrometrije(ICP-OES). Za izdvajanje ekstrakata je korištena klasičnaekstrakciona tehnika prema Soxhlet-u uz upotrebu 80 %etanola (v/v). Spektrofotometrijskim metodama određenje sadržaj ukupnih polifenola, flavonida, flavonola,ukupnih i monomernih antocijana. Kvalitativnom ikvantitativnom HPLC metodom je utvrđen sadržajpojedinačnih polifenolih jedinjenja. Biološki potencijaluzoraka je utvrđen u sistemima in vitro, gdje je određena:antioksidativna aktivnost, antihiperglikemijska aktivnost,antiproliferativni efekat i antimikrobna aktivnost.Antioksidativna aktivnost uzoraka je ispitana sa četirimetode: DPPH test, ABTS test, sposobnost neutralizacijehidroksil radikala i metodom inhibicije Briggs Rauscherovih oscilatornih reakcija. Antihiperglikemijska aktivnostispitivanih uzoraka kupine je dokazan na osnovusposobnosti inhibicije enzima α-glukozidaze.Antiproliferativni efekat ispitivanih uzoraka je određenprema inhibiciji rasta četiri humane ćelijske linije: HeLa(epitelnog karcinoma cerviksa), HT-29 (adenokarcinomadebelog crijeva), MRC-5 (zdravim ćelijama fibroblastipluća) i prema ćelijskoj liniji MCF7 (adenokarcinomadojke). Antimikrobna aktivnost je utvrđena prema grampozitivnom soju (G⁺) bakterija S. Aureus i gramnegativnom (G^-) soju E. coli, rastu micelija Aspergillusaniger i Candida albicans. Regresionom analizom premaPirson-u su određeni odnosi između sadržaja polifenolnihjedinjenja i biološkog potencijala, sa statističkimznačajem (p ≤ 0,01).
The chemical, mineral, polyphenolics composition andbiological potentials of four blackberries varieties from twodifferent locations in the northwestern part of Bosnia andHerzegovina (Verići - wild and cultivated variety Čačanskabestrna and Javorani - wild and cultivated variety ChesterThornless) were determined. The contents of dry matter, ash,crude cellulose, total sugars, total acidity and ascorbic acid wereobtained. The contents of mineral matter were detected byoptical emission spectrometry (ICP-OES) method. A classicSoxhlet extraction technique with 80% ethanol (v/v) wasapplied to obtain extracts. The spectrophotometric methodswere used to determine the content of total polyphenolics,flavonoids, flavonols, and total and monomer anthocyanins. Thecontent of individual polyphenolic compounds was determinedby HPLC method. The biological potentials (antioxidantactivity, antihyperglycemic activity, antiproliferative effect andantimicrobial activity) of the samples were determined in vitrosystems. The antioxidant activity was tested with four methods:DPPH test, ABTS test, ability to neutralize OH radicals and themethod of inhibiting Briggs Rauscher oscillatory reactions. Theantihyperglycaemic activity of the tested blackberry sampleswas based on the ability to inhibit α-glucosidase enzyme. Theantiproliferative effect of the tested samples was determined byinhibiting the growth of four human cell lines: epithelialcarcinoma of the cervix (HeLa), colon adenocarcinoma (HT-29), healthy lung fibroblast cells (MRC-5) and the cell line ofbreast adenocarcinoma (MCF7). The extracts showed thehighest inhibitory effect on the cell line of breastadenocarcinoma (MCF-7). The antimicrobial activity wasdetermined according to gram-positive bacteria (G ⁺) ofStaphylococcus aureus, Gram-negative bacteria (G ^-)Escherichia coli, growth of mycelium Aspergillus niger andfungi Candida albicans. The Pearson correlations werestatistically determined the relationship between the content ofpolyphenolic compounds and biological potential, withstatistical significance (p ≤ 0.01).

L'apprentissage automatique a été utilisé pour diverses tâches d'accès à l'information, tels que la catégorisation, le clustering ou l'extraction d'information. Acquérir les données annotées nécessaires pour appliquer les techniques d'apprentissage supervisé est un défi majeur pour cesapplications, en particulier pour les très grandes collections. Au coursdes dernières années, deux grandes approches ont été explorées dans cesens, l'apprentissage semi-supervisé et l'apprentissage actif. Les deuxparadigmes abordent la question du coût d'annotation, mais de deux pointsde vue différents. D'une part, apprentissage semi-supervisé essaied'apprendre en tenant compte à la fois des données annotées etnon-annotées. D'autre part, l'apprentissage actif tente de trouver lesmeilleurs exemples à annoter, afin de réduire au minimum le nombre d'exemples annotés nécessaire. Dans ce travail, nous étudions des extensions de modèles d'aspect pour la tâche de la classification, où les données sont partiellement annotées.

У агроеколошким условима Западне Србије, у периоду од 2008. до 2010. године,испитиване су помолошке особине сорти високожбунасте боровнице у циљу препоруке за увођење у производњу ове, у светским оквирима, све траженије врстевоћака.Оглед је постављен на објекту „Чачак”, Института за воћарство у Чачку са по 15 биљака сората ‘Bluecrop’ (контрола), ’Duke’, ‘Reka’, ‘Nui’ и ‘Ozarkblu’. Проучаване су физиолошке особине (фенофаза листања, фенофаза цветања, фенофаза опрашивања и оплођења, фенофаза зрења плода, фенофаза отпадања лишћа и дужина вегетационог периода), док су од помолошких особина испитиване физичке особинеплодова и принос. Квалитет плода је оцењиван на основу хемијског састава и органолептичке оцене. Истраживања су, такође, једним делом била усмерена на добијање нових сазнања из области репродуктивне биологије високожбунастеборовнице, односно одвијање прогамне фазе оплођења.Добијени резултати указују да су агроеколошки услови Западне Србије, уз адекватну примену агротехничких и помотехничких мера, погодни за комерцијално гајење испитиваних сорти високожбунасте боровнице. То се посебно односи на сорте ‘Bluecrop’, ‘Duke’ и ‘Ozarkblue’, које су оствариле високе приносе, као и добар квалитет плодова. Захваљујући одличној родности, али због слабијег квалитета плода, сорта ‘Reka’ се може препоручити за гајење у ограниченом обиму,првенствено као опрашивач за сорте ‘Bluecrop’ и ‘Duke’. Без обзира на одличан квалитет плодова, због мањих приноса, ограничени обим гајења препоручује се и за сорту ‘Nui’.
U agroekološkim uslovima Zapadne Srbije, u periodu od 2008. do 2010. godine,ispitivane su pomološke osobine sorti visokožbunaste borovnice u cilju preporuke za uvođenje u proizvodnju ove, u svetskim okvirima, sve traženije vrstevoćaka.Ogled je postavljen na objektu „Čačak”, Instituta za voćarstvo u Čačku sa po 15 biljaka sorata ‘Bluecrop’ (kontrola), ’Duke’, ‘Reka’, ‘Nui’ i ‘Ozarkblu’. Proučavane su fiziološke osobine (fenofaza listanja, fenofaza cvetanja, fenofaza oprašivanja i oplođenja, fenofaza zrenja ploda, fenofaza otpadanja lišća i dužina vegetacionog perioda), dok su od pomoloških osobina ispitivane fizičke osobineplodova i prinos. Kvalitet ploda je ocenjivan na osnovu hemijskog sastava i organoleptičke ocene. Istraživanja su, takođe, jednim delom bila usmerena na dobijanje novih saznanja iz oblasti reproduktivne biologije visokožbunasteborovnice, odnosno odvijanje progamne faze oplođenja.Dobijeni rezultati ukazuju da su agroekološki uslovi Zapadne Srbije, uz adekvatnu primenu agrotehničkih i pomotehničkih mera, pogodni za komercijalno gajenje ispitivanih sorti visokožbunaste borovnice. To se posebno odnosi na sorte ‘Bluecrop’, ‘Duke’ i ‘Ozarkblue’, koje su ostvarile visoke prinose, kao i dobar kvalitet plodova. Zahvaljujući odličnoj rodnosti, ali zbog slabijeg kvaliteta ploda, sorta ‘Reka’ se može preporučiti za gajenje u ograničenom obimu,prvenstveno kao oprašivač za sorte ‘Bluecrop’ i ‘Duke’. Bez obzira na odličan kvalitet plodova, zbog manjih prinosa, ograničeni obim gajenja preporučuje se i za sortu ‘Nui’.

Tato práce se zaměřuje na implementaci počítačového systému rozpoznávání řečníka do prostředí mobilního telefonu. Je zde popsán princip, funkce, a implementace rozpoznávače na mobilním telefonu Nokia N900.

This PhD research has provided novel solutions to three major challenges which have prevented the wide spread deployment of speaker recognition technology: (1) combating enrolment/ verification mismatch, (2) reducing the large amount of development and training data that is required and (3) reducing the duration of speech required to verify a speaker. A range of applications of speaker recognition technology from forensics in criminal investigations to secure access in banking will benefit from the research outcomes.

In this paper, we propose techniques for adaptation of speaker recognition systems. The aim of this work is to create adaptation for Probabilistic Linear Discriminant Analysis. Special attention is given to unsupervised adaptation. Our test shows appropriate clustering techniques for speaker estimation of the identity and estimation of the number of speakers in adaptation dataset. For the test, we are using NIST and Switchboard corpora.

U radu su prikazani eksperimentalni rezultati natkritične ekstrakcije etarskog ulja bobicakleke (Juniperus communis L.) i ulja iz semena uljane tikve golice (Cucurbita pepo L.). Ispitan je uticaj pritiska, temperature, stepena usitnjenosti čestica i protoka natkritičnogugljendioksida na promenu prinosa ulja sa vremenom. U cilju poređenja, usitnjeno seme uljane tikve ekstrahovano je i u ekstraktoru većih dimenzija, NOVA-SWISS, Highpressure extraction plant, kao i heksanom i petroletrom u ekstraktoru tipa Sokslet.Praćena je i promena kvaliteta ekstrakata sa vremenom: u etarskom ulju kleke, GC-FID i GC-MS metodama, određen je relativni sadržaj 50 terpenskih jedinjenja i sve komponente ulja su svrstane u 5 osnovnih grupa (monoterpene, seskviterpene, oksidovane monoterpene, oksidovane seskviterpene i ostale komponente). U tikvinom ulju ekstrahovanom natkritičnim ugljendioksidom određen je masnokiselinski sastav GC-MS analizom, sadržaj tokoferola HPLC analizom, sterola i skvalena GC-MS metodom. Određeni su uslovi koji favorizuju ekstrakciju ispitanih jedinjenja za obe sirovine. Dat je dateljan prikaz matematičkih modela koji se koriste za opisivanje natkritične ekstrakcije etarskih ulja i masnih ulja, počevši od najopštijeg modela koji uključuje diferencijalne bilanse mase za rastvorak u masi natkritičnog fluida, u fluidu unutar pora čestica usitnjenog matrijala i u čvrstoj fazi, koji se uvođenjem određenih pretpostavki pojednostavljuje i svodi na modele koji su izabrani da budu ispitani u okviru ovoga rada. Ispitani su modeli kreireni po analogiji sa hlađenjem vrele kugle u masi fluida, tj. modeli tipa jedne sfere i to: Model jedne sfere-1 (MJS-1), koji pored uticaja koeficijenta efektivne difuzije ulja u materijalu na brzinu prenosa mase uzima u obzir uticaj koeficijenta prenosa mase kroz film natkritičnog fluida oko čestice, pri čemu je njegova vrednost procenjena preko postojećih korelacija; MJS-1 (2 par), u kome je spoljašnji koeficijent prenosa mase uzet kao drugi prilagodljiv parametar modela; MJS-2, gde je koeficijent efektivne difuzije jedini prilagođeni parametar, Model karakterističnog vremena i prošireni model klipnog toka koji je predložila Sovová. Za modelovanje natkritične ekstrakcije ulja semena tikve korišćen je i kombinovani model Honga i sar. Softverskim paketima Mathcad 2001 Professional i Solver dodatka unutar Microsoft Excel 2003, određeni su parametri ispitanih modela u cilju najboljeg slaganja modela sa ekperimentalnim podacima. Za obe ispitane sirovine, među ispitanim modelima, izabrani su modeli koji najbolje opisuju njihovu ekstrakciju natkritičnim ugljendioksidom. Prošireni model klipnog toka koji je predložila Sovová pokazao se podjednako dobrim za modelovanje natkritične ekstrakcije obe sirovine i nešto bolji od ostalih primenjenih modela.
This study provides results of supercritical carbon dioxide (SCCO₂) extraction of juniper berries (Juniperus communis L.) and pumpkin seeds (Cucurbita pepo L. convar. citrullina) in a laboratorysupercritical fluid extraction apparatus. The influenceof pressure, temperature, particle size and carbon dioxide flow on the extraction kinetics of pumpkin seed oil and juniper berry essential oil was studied. Ground pumpkin seeds were also extracted with supercritical carbon dioxide in NOVA-SWISS, High Pressure Extraction Plant, and with hexane and petroleum ether in a laboratory Soxhlet extractor. This work was also aimed to investigate the evolution of the composition of juniper fruit supercritical CO₂ extracts with time, at different extraction pressures and to emphasize the most favorable condition for the extraction of different terpene hydrocarbon groups, reporting the qualitative differences among extracts collected during successive extraction time periods. Juniper berry extracts were analyzed by capillary gas chromatography, using flame ionization (GC-FID) and mass spectrometric detection (GC-MS). More than 200 constituents were detected in the extracts and the contents of 50 compounds were reported in the work. Dependence of the percentage yields of monoterpene, sesquiterpene, oxygenated monoterpene and oxygenated sesquiterpene hydrocarbon groups on extraction time was investigated and conditions that favored the yielding of each terpene groups were emphasized. GC-MS analysis of FAME, prepared by transesterification of pumpkin seed oil with KOH in methanol, was performed. Fatty acid compositions of supercritical CO₂ pumpkin seed extract fractions collected in successive time intervals over the course of the extraction were determined. The same fractions were analyzed by high pressure liquid chromatography (HPLC), using diode-array detector (DAD) in order to determine a- and g-tocopherol contents. Sterol and squalene contents were determined by GC-MS analysis, as well. Conditions that favored the yielding of tocopherols, squalene and sterols were emphasized. A general mass transfer model and its simlifications were analysed. Extraction curves were evaluated by “hot sphere” mathematical models SSM-1 (Single Sphere Model 1 – in which the external mass transfer coefficient also influences the extraction profile and film mass transfer coefficients were estimated by the correlations), SSM-1 (2 par) (film mass transfer coefficient is used as the second adjustable parameter), SSM-2 (only effective diffusivity influence is considered), Characteristic time model and by the extended Lack’s plug-flow model given by Sovová. A combined model of Hong et al. was also fitted to the experimental data for pumpkin seed oil SCCO₂ extractions. Relative merits of the models are demonstrated. Good agreement between the extended Lack’s plug-flow model and the experimental measurements was obtained.

L'increment de bases de dades que cada vegada contenen imatges més difícils i amb un nombre més elevat de categories, està forçant el desenvolupament de tècniques de representació d'imatges que siguin discriminatives quan es vol treballar amb múltiples classes i d'algorismes que siguin eficients en l'aprenentatge i classificació. Aquesta tesi explora el problema de classificar les imatges segons l'objecte que contenen quan es disposa d'un gran nombre de categories. Primerament s'investiga com un sistema híbrid format per un model generatiu i un model discriminatiu pot beneficiar la tasca de classificació d'imatges on el nivell d'anotació humà sigui mínim. Per aquesta tasca introduïm un nou vocabulari utilitzant una representació densa de descriptors color-SIFT, i desprès s'investiga com els diferents paràmetres afecten la classificació final. Tot seguit es proposa un mètode par tal d'incorporar informació espacial amb el sistema híbrid, mostrant que la informació de context es de gran ajuda per la classificació d'imatges. Desprès introduïm un nou descriptor de forma que representa la imatge segons la seva forma local i la seva forma espacial, tot junt amb un kernel que incorpora aquesta informació espacial en forma piramidal. La forma es representada per un vector compacte obtenint un descriptor molt adequat per ésser utilitzat amb algorismes d'aprenentatge amb kernels. Els experiments realitzats postren que aquesta informació de forma te uns resultats semblants (i a vegades millors) als descriptors basats en aparença. També s'investiga com diferents característiques es poden combinar per ésser utilitzades en la classificació d'imatges i es mostra com el descriptor de forma proposat juntament amb un descriptor d'aparença millora substancialment la classificació. Finalment es descriu un algoritme que detecta les regions d'interès automàticament durant l'entrenament i la classificació. Això proporciona un mètode per inhibir el fons de la imatge i afegeix invariança a la posició dels objectes dins les imatges. S'ensenya que la forma i l'aparença sobre aquesta regió d'interès i utilitzant els classificadors random forests millora la classificació i el temps computacional. Es comparen els postres resultats amb resultats de la literatura utilitzant les mateixes bases de dades que els autors Aixa com els mateixos protocols d'aprenentatge i classificació. Es veu com totes les innovacions introduïdes incrementen la classificació final de les imatges.
The release of challenging data sets with ever increasing numbers of object categories is
forcing the development of image representations that can cope with multiple classes and
of algorithms that are efficient in training and testing. This thesis explores the problem of
classifying images by the object they contain in the case of a large number of categories. We first investigate weather the hybrid combination of a latent generative model with a discriminative classifier is beneficial for the task of weakly supervised image classification.
We introduce a novel vocabulary using dense color SIFT descriptors, and then investigate classification performances by optimizing different parameters. A new way to incorporate spatial information within the hybrid system is also proposed showing that contextual information provides a strong support for image classification. We then introduce a new shape descriptor that represents local image shape and its spatial layout, together with a spatial pyramid kernel. Shape is represented as a compact
vector descriptor suitable for use in standard learning algorithms with kernels. Experimental
results show that shape information has similar classification performances and sometimes outperforms those methods using only appearance information. We also investigate how different cues of image information can be used together. We
will see that shape and appearance kernels may be combined and that additional information
cues increase classification performance. Finally we provide an algorithm to automatically select the regions of interest in training. This provides a method of inhibiting background clutter and adding invariance to the object instance's position. We show that shape and appearance representation over the regions of interest together with a random forest classifier which automatically selects the best cues increases on performance and speed. We compare our classification performance to that of previous methods using the authors'own datasets and testing protocols. We will see that the set of innovations introduced here lead for an impressive increase on performance.

碩士
國立雲林科技大學
資訊管理系碩士班
99
The population of the Internet and digital device made a lot of images existing in the Internet without arrangement. Thus, how to organize these images and to make them easy to find is an important issue. One way to solve this issue is automatic image annotation (AIA). This method can save manpower to manage images and to make image annotations consistence. But the annotation would be limited by already defined terms. Besides, synonymous and homonym are also existing between terms. In order to make extra meaningful terms in web images annotation, and to modify synonymous and homonym problems, this study using probabilistic latent semantic analysis (PLSA)to analyze amount of web pages to gather parameters and to make annotation on web pages. This study also compares the number of keywords gathering from news, to weight on keywords for increasing annotation accuracy. The average precision is 85%, recall is 71%.

碩士
輔仁大學
資訊工程學系碩士班
102
The aim of this thesis is to category the movies into genres using the previews. This study attempts to combine audio, visual and text features to classify a collection of movie previews into action, biography, comedy, and horror. For each of the collected previews, the audio and visual features are extracted and the text features are drawn from social tags via social websites. The probabilistic latent semantic analysis (PLSA) is used to incorporate the features from these three different aspects of information.The standard PLSA processes one type of information only.Therefore double-modeland triple-model PLSAsare extended to combinetwo or three different types of information. We compare these various variants of PLSA approaches with unimodal PLSAs, which use either audio, visual or text features only. The experimental results show that one of triple-model PLSAs achievesthe highest accuracy, and social tags (text features) play an important role for classifying movies genres.

碩士
國立交通大學
多媒體工程研究所
100
Document classification is of great practical importance today given the massive volume of online text available. Supervised learning is one of the popular techniques for tackling document classification problems. However, sufficient labeled data is necessary for supervised learning methods to train a classification model. Labeling must typically be done manually and it is a time-consuming process obviously. In general, unlabeled data may be relatively easy to collect. Although unsupervised learning methods don’t need any labeled data, users often have some background knowledge before clustering. Practically, background knowledge should be considered in the algorithms to improve clustering accuracy. This paper proposes a semi-supervised learning algorithm, which considers dimension reduction and clustering simultaneously. This paper applies constrained-pLSA to obtain soft labels , and then combines soft labels with linear discriminant analysis to find a better feature space. We conduct experiments on CiteUlike, 20Newsgroups, Reuters and experimental results indicate that the proposed method can effectively improve clustering performance.

碩士
輔仁大學
資訊工程學系
101
In web 2.0 era, social tags have become a powerful indirect information in the interpretation of musical semantics and received growing interests in music information retrieval. This paper proposes an approach based on probabilistic latent semantic analysis (PLSA) model for music genre classification. We describe and derive three models: (1) a double-layer tag-based PLSA, (2) a simple multimodal PLSA that integrates two single-layer PLSAs, one is tag-based and the other is audio content-based, and (3) an elaborated multimodal PLSA that integrates a double-layer tag-based PLSA with a single-layer audio content-based PLSA. Experimental results based on SVM classifier demonstrate the effectiveness of our models. In particular, the results indicate the elaborated multimodal PLSA is the best proposed model for music genre classification.

碩士
國立臺灣海洋大學
資訊工程學系
103
Object classification is an important research of the image processing. In many papers, they use many ways to classify the objects, like Sparse Representation Classification (SRC). SRC needs to calculate the residual of reconstruction error and to find the best candidate. It is very inefficient because of an optimization process is involved. In addition, it uses only the residual that ignore to consider the distribution of combination coefficients of visual codes in classification. Thus, it often fails to classify categories when they are similar. In this thesis, we proposed a novel classification method that combined probabilistic latent semantic analysis (pLSA) and sparse representation. We use the results of vehicle color classification experimental to prove our proposed method is to improve the SRC.

碩士
國立臺灣海洋大學
資訊工程學系
101
This thesis proposes a novel object classification method combined sparse coding (SR) and probabilistic latent semantic analysis (pLSA) with a new classifier and adaptive kNN. Sparse coding is widely used in image processing. Many papers applied sparse coding in image classification, most of which chose reconstruction error or improved methods as their classifier. Although these kinds of methods are precise, the restraint (Sparsity and Redundancy) make it performs not fine under low dimension feature and small training dataset. This thesis solved the problem by using a new classifier based on pLSA. Nowadays, algorithms in natural language processing like LSA and pLSA are widely used in image processing. Derived from the idea of Bag-of-Words, pLSA-based classifiers have been applied in many papers. In this thesis, we intend a new classifier which both achieves high accurate rate and high efficiency. A new adaptive kNN algorithm is also proposed in our research to improve the correct rate of original kNN and reduce the time cost in weighted kNN. A novel real-time object recognition system is framed based on the given combined method. The proposed approach repeatedly achieves state-of-the-art results on private datasets and several public data sets.

Les tourbières ombrotrophes, ou bogs sont particulièrement vulnérables à l’augmentation de la déposition atmosphérique d’azote. Cet apport d’un nutriment normalement limitant altère la capacité des tourbières à accumuler le carbone (C), en plus de mener à des changements de leur composition végétale. L’imagerie spectrale est une approche prometteuse puisqu’elle rend possible la détection des espèces végétales et de certaines caractéristiques chimiques des plantes, à distance. Toutefois, l’ampleur des différences spectrales intra- et interespèces n’est pas encore connue. Nous avons évalué la façon dont la chimie, la structure et la signature spectrale des feuilles changent chez Chamaedaphne calyculata, Kalmia angustifolia, Rhododendron groenlandicum et Eriophorum vaginatum, dans trois tourbières du sud du Québec et de l’Ontario, incluant une tourbière où se déroule une expérience de fertilisation à long terme. Nous avons mesuré des changements dans les traits fonctionnels dus aux différences dans la quantité d’azote disponible dans les sites. Toutefois, la déposition atmosphérique d’azote a eu relativement peu d’effet sur les spectres foliaires ; les variations spectrales les plus importantes étaient entre les espèces. En fait, nous avons trouvé que les quatre espèces ont un spectre caractéristique, une signature spectrale permettant leur identification au moyen d’analyses discriminantes des moindres carrés partiels (PLSDA). De plus, nous avons réussi à prédire plusieurs traits fonctionnels (l’azote, le carbone ; et la proportion d’eau et de matière sèche) avec moins de 10 % d’erreur grâce à des régressions des moindres carrés partiels (PLSR) des données spectrales. Notre étude fournit de nouvelles preuves que les variations intraspécifiques, causées en partie par des variations environnementales considérables, sont perceptibles dans les spectres foliaires. Toutefois, les variations intraspécifiques n’affectent pas l’identification des espèces ou la prédiction des traits. Nous démontrons que les spectres foliaires comprennent des informations sur les espèces et leurs traits fonctionnels, confirmant le potentiel de la spectroscopie pour le suivi des tourbières.
Abstract Bogs, as nutrient-poor ecosystems, are particularly sensitive to atmospheric nitrogen (N) deposition. Nitrogen deposition alters bog plant community composition and can limit their ability to sequester carbon (C). Spectroscopy is a promising approach for studying how N deposition affects bogs because of its ability to remotely determine changes in plant species composition in the long term as well as shorter-term changes in foliar chemistry. However, there is limited knowledge on the extent to which bog plants differ in their foliar spectral properties, how N deposition might affect those properties, and whether subtle inter- or intraspecific changes in foliar traits can be spectrally detected. Using an integrating sphere fitted to a field spectrometer, we measured spectral properties of leaves from the four most common vascular plant species (Chamaedaphne calyculata, Kalmia angustifolia, Rhododendron groenlandicum and Eriophorum vaginatum) in three bogs in southern Québec and Ontario, Canada, exposed to different atmospheric N deposition levels, including one subjected to a 18 years N fertilization experiment. We also measured chemical and morphological properties of those leaves. We found detectable intraspecific changes in leaf structural traits and chemistry (namely chlorophyll b and N concentrations) with increasing N deposition and identified spectral regions that helped distinguish the site-specific populations within each species. Most of the variation in leaf spectral, chemical and morphological properties was among species. As such, species had distinct spectral foliar signatures, allowing us to identify them with high accuracy with partial least squares discriminant analyses (PLSDA). Predictions of foliar traits from spectra using partial least squares regression (PLSR) were generally accurate, particularly for the concentrations of N and C, soluble C, leaf water, and dry matter content (<10% RMSEP). However, these multi-species PLSR models were not accurate within species, where the range of values was narrow. To improve the detection of short-term intraspecific changes in functional traits, models should be trained with more species-specific data. Our field study showing clear differences in foliar spectra and traits among species, and some within-species differences due to N deposition, suggest that spectroscopy is a promising approach for assessing long-term vegetation changes in bogs subject to atmospheric pollution.

碩士
國立交通大學
資訊科學與工程研究所
99
Text classification is of great practical importance today given the massive volume of online text available. Supervised learning is one of the popular techniques for tackling text classification problems. However, enough labeled data is necessary for supervised learning methods. Labeling must typically be done manually and it is a time-consuming process obviously. In general, unlabeled data may be relatively easy to collect. Although unsupervised learning method doesn’t need any labeled data. But users often have some background knowledge before clustering. Practically, background knowledge should be included into algorithms to improve clustering accuracy. This paper extends PLSA clustering model to propose a Constrained-PLSA method, which is a semi-supervised learning algorithm. The Constrained-PLSA assumes that data is generated by a mixture model and the correspondence between each document and class label is one to one. By introducing the seeding documents as constraints, we show that Constrained-PLSA can estimate maximum likelihood in latent variable models using the Expectation Maximization (EM) algorithm. Experimental results show that Constrained-PLSA with a small amount of examples can effectively improve the performance. In addition, this paper also discusses tag usage using Constrained-PLSA. Academic paper data set is employed in this paper. Each paper consists of abstract and tag information. Tag is given by users after reading the article. This paper analyzes four combinations of abstracts and tags: “words only”, “tags only”, “words + tags” and “tags as words”. The best one is presented in this paper. Meanwhile, the experimental result shows that Constrained-PLSA outperforms other clustering algorithms.

Αντικείμενο της παρούσας εργασία είναι ο υπολογισμός και η ταξινόμηση, με βάση το σύστημα, BIRADS της πυκνότητας του μαστού από εικόνες μαστογραφίας. Στα πλαίσια της προσπάθειας αυτής αναπτύχθηκε ολοκληρωμένο υπολογιστικό σύστημα σε γραφικό περιβάλλον ως λογισμικό πακέτο, σε γλώσσα Visual C++ .NET . Το υπολογιστικό αυτό σύστημα δέχεται σαν είσοδο εικόνες μαστογραφίας σε οποιοδήποτε από τα δημοφιλή bitmap format εικόνων όπως jpeg και tiff καθώς και DICOM αρχεία. Η λειτουργία του μπορεί να χωριστεί σε τρία στάδια: το στάδιο της προεπεξεργασίας, το στάδιο απομόνωσης της περιοχής του μαστού και το στάδιο καθορισμού της πυκνότητας του μαστού. Στο πρώτο στάδιο παρέχονται μια σειρά από στοιχειώδη εργαλεία επεξεργασίας εικόνας όπως εργαλεία περιστροφής, αποκοπής και αλλαγής αντίθεσης . Επιπρόσθετα παρέχεται η δυνατότητα Ανισοτροπικού Φιλτραρίσματος της εικόνας. Στο δεύτερο στάδιο γίνεται η απομόνωση της περιοχής του μαστού είτε απευθείας από τον χρήστη είτε αυτόματα με χρήση των ιδιοτήτων του μονογονικού (monogenic) σήματος για την αφαίρεση του παρασκηνίου (background) καθώς και κυματιδίων Gabor για τον διαχωρισμού του θωρακικού μυός. Στο τρίτο στάδιο παρέχεται η δυνατότητα ταξινόμησης της πυκνότητας του μαστού από τον χρήστη με τον καθορισμό κατάλληλου κατωφλίου των επιπέδων γκρίζου της εικόνας αλλά και η δυνατότητα αυτόματης ταξινόμησης της πυκνότητας του μαστού κατά BIRADS με χρήση Δομικών Στοιχείων Υφής (textons) και της τεχνικής pLSA. Όλες οι παραπάνω λειτουργίες παρέχονται μέσω μίας κατά το δυνατόν φιλικότερης προς τον χρήστη διεπαφής.
The present thesis aims at the classification of breast tissue according to BIRADS system based on texture features. To this end an integrated software system was developed in visual C ++. The system takes as inputs pictures in most of the popular bitmap formats like .jpeg and .till as well as DICOM. The functionality of the system is provided by three modules: (a) pre-processing module, (b) breast segmentation module and (c) the breast tissue density classification module. In the pre-processing module a set tools for image manipulation (rotation, crop, gray level adjustment) are available which are accompanied by the ability to perform anisotropic filtering to the input image. In the second module, the user has the ability to interactively define the actual borders of the breast or ask the system to perform it automatically. Automatic segmentation is a two step procedure; in the first step breast tissue is separated from its background by using the characteristics of monogenic signals, while in the second step the pectoral muscle region is subtracted using Gabor wavelets. In the density classification module the user can either ask for a calculation of breast density based on user-defined grey level threshold or perform an automatic BIRADS-based classification using texture characteristics in conjunction with Probabilistic Latent Semantic Analysis (pLSA) algorithm. Special emphasis was given to the development of a functional and user-friendly interface.