Heating milk to high temperatures such as 140 ºC, as used in ultra high temperature (UHT) processing, causes physical and chemical changes in the milk. The production of a cooked flavour is a major change which reduces consumer acceptance of the UHT milk. It has been correlated with the formation of volatile sulphur compounds (VSCs) that result from milk proteins, principally the whey proteins β-lactoglobulin, containing the the sulphur amino acids cystine, cysteine and methionine. The VSCs in milk, whose concentrations are in the parts per billion to parts per million range, are highly reactive, easily oxidised, and sensitive to heat during thermal processing and analysis; this makes them a challenge to analyse. A sensitive method based on gas chromatography with pulsed flame photometric detection coupled with headspace sampling by solid phase microextraction (SPME/GC/PFPD) was developed to detect these compounds in commercial UHT milk and to investigate their production and disappearance during heating and storage. The SPME/GC/PFPD procedure was optimised using different extraction time (15 min, 30 min, & 60 min) – temperature (30 oC, 45 oC & 60 oC) combinations with CAR/PDMS fibre to obtain maximum sensitivity. A short extraction time (15 min) at low temperature (30 oC) was chosen to provide high sensitivity for detecting all VSCs in UHT milk without introducing artefactual VSCs. The extraction method and GC run time (16 min) make this method simple and fast. Nine VSCs were detected in commercial indirectly processed UHT milk, skim and whole. These are hydrogen sulphide (H2S), carbonyl sulphide (COS), methanethiol (MeSH), dimethyl sulphide (DMS), carbon disulphide (CS2), dimethyl disulphide (DMDS), dimethyl sulphoxide (DMSO), dimethyl sulphone (Me2SO2) and dimethyl trisulphide (DMTS). An additional unknown compound was detected but could not be identified by GC/MS because its concentration was below the detection limit of the MS detector. The concentrations of H2S, DMS and DMTS were higher than their threshold values indicating their importance in milk flavour, especially cooked flavour. Several attempts have been made to reduce the cooked flavour in UHT milk. In the current research, the use of hydrogen peroxide (H2O2) to oxidise the VSCs and thereby reduce cooked flavour was investigated. H2O2 is used as a milk preservative and is generally recognised as safe (GRAS) in USA. Several concentrations of H2O2 (0.001%, 0.005%, 0.01%, 0.02% & 0.03%) were added to milk to assess its effects on VSCs and on whey proteins denaturation in UHT milk. H2O2 effectively reduced the concentration of all VSCs, except DMDS which was increased, presumably by oxidation of MeSH. H2S was completely oxidised or reduced below its threshold value. Low concentrations of H2O2 (0.001% & 0.005%) had no effect on, or decreased, the extent of denaturation of β-lactoglobulin when added after or before processing, respectively. Some UHT plants use severe heating conditions, leading to high levels of denaturation of whey proteins, particularly β-Lg, the main source of the VSCs in milk. Correlations between heat severity, β-Lg denaturation and individual VSC generation were investigated in milk batch-heated at 80 oC and 90 oC, and UHT milk processed at 120-150 oC. In accordance with previous reports, β-Lg was more heat-sensitive than α-La. Only five VSCs were detected. The concentrations of H2S and MeSH correlated well with denaturation of β-Lg and α-La. DMS concentration correlated well with β-Lg in UHT milk but not in the batch-heated milk. CS2 did not show a good correlation with heat intensity and appeared to plateau out after a certain level of heating. Conversely, COS and MeSH seemed to require a certain minimum amount of heat before generation commenced; this corresponded to denaturation of β-Lg above 49% and 89% respectively at 80 oC. The higher concentrations of DMS and H2S in UHT milk compared with batch-heated samples having similar degrees of denaturation suggested other possible sources for their production and the importance of the heat severity in generating them. For example, at high heat intensity, S-methylmethionine and thiamine could be sources of DMS and H2S respectively. Furthermore, in whole milk as used in this work, milk fat globule membrane proteins are another source of VSCs. The outcome of this study will help UHT manufacturers to understand the production and disappearance of the VSCs in commercial UHT milk and how to adjust the processing conditions to avoid generation of cooked flavour. Additionally, the promising results of using low concentrations of H2O2 to oxidise the VSCs will provide the industry with another means of reducing cooked flavour. Before H2O2 use is implemented in UHT processing, future studies are required to evaluate all of its effects, including sporicidal effects. Overall, this study makes a contribution to finding a solution to the cooked flavour problem in UHT milk, thereby increasing market share of this milk in countries such as Australia, the UK and North America where cooked flavour is the main barrier to its consumer acceptance.

Orientador: Geraldo de Freitas Maciel / Resumo: Esta tese de doutorado traz uma abordagem experimental sobre mecanismos de geração, desenvolvimento e propagação em canal de instabilidade na superfície livre de um escoamento de fluido não-Newtoniano. Estas instabilidades, quando evoluem para um padrão estável, exibem comprimento, amplitude e celeridade bem definidos, sendo denominadas roll waves. Na literatura existe uma lacuna no que diz respeito às medições destes fenômenos em condições controladas de laboratório, e com pouquíssimos registros, quando presentes em eventos naturais. Assim, buscou-se neste trabalho projetar e implementar um aparato experimental operacional e de baixo custo com o objetivo de gerar e aferir tais instabilidades, sob condições controladas e sem a influência de vibrações externas (canalete posicionado sobre uma mesa inercial), seguindo metodologia de ensaio proposta e testada. O fluido teste utilizado foi o gel de carbopol 996, que apresentou boa estabilidade e propriedades reológicas aderentes ao modelo de Herschel-Bulkley, tal qual as lamas encontradas em pés de barragens, lamas oriundas de fluxos hiperconcentrados ou de corridas propriamente ditas. Um sistema de sucção-recalque, através de uma bomba hidráulica de cavidade progressiva, impulsionava o gel para o canal de ensaio, garantindo fluxo contínuo do material em regime de recirculação. Para gerar as instabilidades na superfície livre, foi imposta, a montante do canal, uma perturbação, por meio de um pulso de ar controlado em intensidade e... (Resumo completo, clicar acesso eletrônico abaixo) / Doutor

Submitted by EVANDRO FERNANDES DA CUNHA (evandro.cunha@unesp.br) on 2018-11-18T20:13:14Z No. of bitstreams: 1 Tese Evandro Cunha.pdf: 10054124 bytes, checksum: f324825069efc486e2637a46c1b85371 (MD5) / Approved for entry into archive by Cristina Alexandra de Godoy null (cristina@adm.feis.unesp.br) on 2018-11-21T15:21:55Z (GMT) No. of bitstreams: 1 cunha_ef_dr_ilha.pdf: 10054124 bytes, checksum: f324825069efc486e2637a46c1b85371 (MD5) / Made available in DSpace on 2018-11-21T15:21:55Z (GMT). No. of bitstreams: 1 cunha_ef_dr_ilha.pdf: 10054124 bytes, checksum: f324825069efc486e2637a46c1b85371 (MD5) Previous issue date: 2018-07-31 / Esta tese de doutorado traz uma abordagem experimental sobre mecanismos de geração, desenvolvimento e propagação em canal de instabilidade na superfície livre de um escoamento de fluido não-Newtoniano. Estas instabilidades, quando evoluem para um padrão estável, exibem comprimento, amplitude e celeridade bem definidos, sendo denominadas roll waves. Na literatura existe uma lacuna no que diz respeito às medições destes fenômenos em condições controladas de laboratório, e com pouquíssimos registros, quando presentes em eventos naturais. Assim, buscou-se neste trabalho projetar e implementar um aparato experimental operacional e de baixo custo com o objetivo de gerar e aferir tais instabilidades, sob condições controladas e sem a influência de vibrações externas (canalete posicionado sobre uma mesa inercial), seguindo metodologia de ensaio proposta e testada. O fluido teste utilizado foi o gel de carbopol 996, que apresentou boa estabilidade e propriedades reológicas aderentes ao modelo de Herschel-Bulkley, tal qual as lamas encontradas em pés de barragens, lamas oriundas de fluxos hiperconcentrados ou de corridas propriamente ditas. Um sistema de sucção-recalque, através de uma bomba hidráulica de cavidade progressiva, impulsionava o gel para o canal de ensaio, garantindo fluxo contínuo do material em regime de recirculação. Para gerar as instabilidades na superfície livre, foi imposta, a montante do canal, uma perturbação, por meio de um pulso de ar controlado em intensidade e frequência. Após evolução e estabilização, a onda (roll wave) era aferida por meio de sistema de medição ultrassônico na configuração de pulso-eco e, seguindo metodologia inovadora proposta nesta tese, por meio de técnica fotométrica em função da absorção luminosa de um corante (azul de metileno) adicionado ao fluido teste em escoamento. Para confirmação da qualidade dos resultados obtidos pela aplicação das técnicas de medição experimental, foram feitas confrontações entre as técnicas experimentais propostas e entre estas e um modelo numérico, desenvolvido pelo grupo de pesquisa. Por fim, foram testados dois estudos de caso de formação e propagação de roll waves no canal, o primeiro visando a obtenção e tratamento das amplitudes em diversos pontos ao longo do canal, e, no segundo, a influência da frequência de perturbação sobre a amplitude da roll wave, todos para cenários experimentais com números de Froude superiores ao Froude mínimo, condição necessária à geração do fenômeno. / This doctoral thesis brings an experimental approach on generation, development and propagation mechanisms of instabilities free-surface flows of a non-Newtonian fluid. When these instabilities evolve to a stable pattern, they display well-defined length, amplitude and celerity, being denominated as roll waves. There is a great gap in literature concerning experimental data. Few experiments are found relating laboratory measurements of these phenomena in controlled conditions; this is also true when it comes to in situ measurements from natural events. Thus, the aim of this project was to design and implement a low cost operational experimental apparatus to generate and measure such instabilities under controlled conditions, isolated from external vibration influences. This project follows a previous methodology developed which was tested and approved. Carbopol gel 996 was used as test fluid, presenting good stability and rheological properties matching to Herschel-Bulkley model, oftenly used for modeling industrial sludge, slurry fluids, natural mud, among others. A hydraulic circuit was installed in which a progressive cavity hydraulic pump pushed the gel to the inlet of the testing channel, ensuring continuous flow of material, capturing the fluid at the channel outlet. A free-surface disturbance was imposed to the flow through a controlled pulse of air (intensity and frequency) on the upstream flow to produce the free-surface instabilities. After evolution and stabilization, the wave (roll wave) was measured through ultrasonic measurement system in the pulse-echo configuration and also through photometric technique using light absorption by a dye (methylene blue) added to the test fluid. Confrontations of the experimental data were made among the proposed experimental techniques and also with numerical simulations developed by the research group in order to assess the quality of the results obtained by the application of the experimental measurement techniques. Lastly, two case studies of formation and propagation of roll waves were tested: the first one aimed to obtain and process the amplitudes in several points along the channel; the second one, the influence of frequency of disturbance over the amplitude of the roll wave was studied. All the experimental scenarios were tested for Froude number over the minimum value necessary for the generation of the phenomenon.

In this thesis a critical assessment of the ages derived using theoretical pre-main-sequence (pre-MS) stellar evolutionary models is presented by comparing the predictions to the low-mass pre-MS population of 14 young star-forming regions (SFRs) in colour-magnitude diagrams (CMDs). Deriving pre-MS ages requires precise distances and estimates of the reddening. Therefore, the main-sequence (MS) members of the SFRs have been used to derive a self-consistent set of statistically robust ages, distances and reddenings with associated uncertainties using a maximum-likelihood fitting statistic and MS evolutionary models. A photometric method (known as the Q-method) for de-reddening individual stars in regions where the extinction is spatially variable has been updated and is presented. The effects of both the model dependency and the SFR composition on these derived parameters are also discussed. The problem of calibrating photometric observations of red pre-MS stars is examined and it is shown that using observations of MS stars to transform the data into a standard photometric system can introduce significant errors in the position of the pre-MS locus in CMD space. Hence, it is crucial that precise photometric studies (especially of pre- MS objects) be carried out in the natural photometric system of the observations. This therefore requires a robust model of the system responses for the instrument used, and thus the calculated responses for the Wide-Field Camera on the Isaac Newton Telescope are presented. These system responses have been tested using standard star observations and have been shown to be a good representation of the photometric system. A benchmark test for the pre-MS evolutionary models is performed by comparing them to a set of well-calibrated CMDs of the Pleiades in the wavelength regime 0.4−2.5 μm. The masses predicted by these models are also tested against dynamical masses using a sample of MS binaries by calculating the system magnitude in a given photometric band- pass. This analysis shows that for Teff ≤ 4000 K the models systematically overestimate the flux by a factor of 2 at 0.5 μm, though this decreases with wavelength, becoming negligible at 2.2 μm. Thus before the pre-MS models are used to derive ages, a recalibration of the models is performed by incorporating an empirical colour-Teff relation and bolometric corrections based on the Ks-band luminosity of Pleiades members, with theoretical corrections for the dependence on the surface gravity (log g). The recalibrated pre-MS model isochrones are used to derive ages from the pre-MS populations of the SFRs. These ages are then compared with the MS derivations, thus providing a powerful diagnostic tool with which to discriminate between the different pre- MS age scales that arise from a much stronger model dependency in the pre-MS regime. The revised ages assigned to each of the 14 SFRs are up to a factor two older than previous derivations, a result with wide-ranging implications, including that circumstellar discs survive longer and that the average Class II lifetime is greater than currently believed.

J'ai consacré mon travail de thèse à la recherche de quasars à haut redshift dans le sondage Canada France High-z Quasar Survey in the Near Infrared (CFHQSIR). L'objectif principal de ce programme est de découvrir des quasars de redshift de l'ordre de z $\sim$ 7 dans les champs larges du CFHTLS (CFHTLS Wide) en utilisant des images réalisées dans le proche infrarouge avec la caméra WIRCam installée au foyer du CFHT. J'ai tout d'abord effectué la calibration photométrique de l'ensemble des données de CFHQSIR. J'ai ensuite étudié la qualité des images ainsi que leur propriété de bruit et leur profondeur. J'ai dans un deuxième temps calculé le nombre de quasars qu'il serait possible de détecter avec CFHQSIR et en ai déduit les contraintes envisageables sur la fonction de luminosité des quasars à z $\sim$ 7. J'ai ensuite procédé à l'identification de candidats quasars parmi les dizaines de milliers de sources que comptent les 130 degrés carrés couvert par CFHQSIR. Des observations de suivi photométriques des candidats ont finalement permis de révéler une cinquantaine d’objets, dont une quinzaine a été sélectionnée pour des observations spectroscopiques au Very Large Telescope (VLT). Celles-ci n’ayant, pour la plupart, pas encore été realisées, la nature exacte de ces sources ne pourra être connue que dans les prochains mois. Pour finir, la dernière partie de ma thèse s’est focalisée sur l’étude de modèles statistiques bayésiens afin de compléter ma méthode de sélection de candidats quasars. Cette étude m'a finalement permis de vérifier que la majorité des objets retenus pour des observations spectroscopiques étaient effectivement les candidats les plus probables. / My PhD work is focused on the search for high-redshift quasars in the Canada France High-z Quasar Survey in the Near Infrared (CFHQSIR). The main scientific objective of this CFHT Large Program is to search for quasars at redshift z $\sim$ 7 with near-infrared images of the CFHTLS Wide fields acquired with the CFHT WIRCam camera. Firstly, I carried out the photometric calibration of the CFHQSIR images. I performed a detailed analysis of the CFHQSIR data by studying their quality, as well as their noise properties and their depths.Secondly, I computed the number of high-redshift quasars that could be detected with CFHQSIR and the corresponding constraints which could be put on the z $\sim$ 7 quasar luminosity function. Then, I proceeded to the identification of quasar candidates among the many thousands of sources in the 130 square degrees covered by CFHQSIR. Photometric follow-up observations of the candidates revealed about fifty objects, of which fifteen were chosen to be spectroscopically observed with the Very Large Telescope (VLT). Given that this spectroscopic follow-up is not yet completed, the exact nature of these sources will only be known in the coming months. To finish, I applied Bayesian model comparison to my sample in order to complete and consolidate my selection procedure. My candidates were finally classified according to their probability to be a high-redshift quasar. I verified that the majority of the most likely candidates were selected for spectroscopic observations.

Les techniques de reconstruction 3d sont devenues incontournables pour des applicationstelles que la caractérisation et l'analyse de surfaces. Les travaux réalisés au coursde cette thèse ont pour objectif d'améliorer la qualité des reconstructions 3d par stéréophotométrie.Cette méthode repose sur deux principes, l'inversion d'un modèle d'interactionlumière/matière (BRDF) et la configuration d'un système d'éclairage et de prises de vues.Pour des surfaces diffuses, la stéréo-photométrie est réalisée à partir d'un minimum detrois images acquises d'un point de vue fixe pour des directions d'éclairages différentes.Son avantage est d'extraire simultanément les propriétés géométriques et colorimétriquesdes surfaces analysées même en cas de forte rugosité. Néanmoins, son application exige laformulation de plusieurs hypothèses qui sont difficilement respectables dans un contexteréel. Ceci génère des erreurs significatives dans les reconstructions. Pour les réduire, nousproposons différentes contributions qui s'articulent autour de la prise en compte globale de lachaine d'acquisition. Les apports de nos travaux se situent aux niveaux de la caractérisationet de la modélisation du système d'éclairage, du capteur d'acquisition et de l'améliorationde la qualité des images. Nous nous sommes aussi intéressés à l'optimisation des protocolesde prises de vues dans le cas de spécularité surfacique ou d'ombrage dus à la présence derugosité. Les résultats obtenus montrent que la prise en compte de ces caractéristiques dansl'inversion d'un modèle de BRDF permet une nette amélioration des reconstructions et offrela possibilité de réduire la taille des systèmes d'acquisition. / Tridimensional reconstruction method has become essential for applications suchas the characterization and analysis of surfaces. In this thesis, aims are to increase the qualityof 3d reconstructions by photometric stereo. This method is based on two principles, reversinglight-matter interaction model and configuration of a lighting system. With diffuse surfaces,the photometric stereo use three captured images from a fixed point of view for differentillumination directions. Its main advantage is to extract the color and geometric propertiesfor the textured rough surfaces. However, its application requires to make assumptions thatare not credible in real cases. This problem generates significant errors in the reconstructions.To reduce them, we offer various solutions around the overall consideration of the acquisitionchain. Our contribution focuses on the characterization and modeling of the lighting system,the acquisition sensor and improved image quality. We are also interested to optimize acquisitionprotocol in the case of specular surface or shading due to the surface geometry. Ourresults show that the inclusion of these features in the inversion of a BRDF model allowsan improvement of 3d reconstructions as well as the possibility of reducing the size of theacquisition systems.

Les lampes à LEDs s’imposent comme la solution d’avenir pour l’éclairage vert de demain. Malgré des durées de vie annoncées par les constructeurs supérieures à 50 000h, une dérive prématurée de la qualité colorimétrique a été constatée. Ce travail de thèse porte sur l’influence de la photodégradation sur les propriétés optiques de l’un des composants d’un module LED : le composite luminescent qui est associé à la puce semi-conductrice. Des composites luminescents ont été élaborés à partir d’une matrice polymère (EVA ou PMMA) et d’un luminophore inorganique : Y3BO6 : 15% Eu3+. L’évolution des propriétés physico-chimiques et optiques de ces composites a été étudiée sous irradiation dans des conditions accélérées (λ>300 nm) et dans les conditions d’usage (LED UV, λ=365 nm). Le luminophore n’a pas montré d’effet prodégradant sur la photooxydation du polymère. Une perte des propriétés optiques du composite sous l’effet de la photodégradation du polymère a été mise en évidence dans le cas de la matrice EVA. L’évolution des paramètres photométriques de la lumière émise par les composites luminescents au cours du photovieillissement a pu être reliée à l’accumulation des photoproduits dans les films. En revanche, pour les composites PMMA/luminophore, la structure chimique et les propriétés optiques de ces composites ne sont pas modifiés pour des durées de photovieillissement accélérés de 4000h. / Phosphor-converted LEDs are emerging as an eco-friendly solution for the next generation lighting. Despite lifetimes claimed over 50 000h, a loss of the optical properties was noticed. This work focuses on the influence of photodegradation of the materials on the optical properties of the polymer/phosphor composite that is used to encapsulate the semiconductor chip. Luminescent composites were made from a polymer matrix (EVA or PMMA) and an inorganic phosphor: Y3BO6 :15% Eu3+. The evolution of physical, chemical and optical properties of these composites was investigated under irradiation in accelerated conditions (λ>300 nm) and in the use conditions of UV LED (λ=365 nm). The phosphor showed no pro-degrading effect on the photooxidation of the polymer. A loss of the optical properties of the composite was observed and ascribed to the photodegradation of the polymer in the case of the EVA matrix. The evolution of photometric parameters of the emitted light by the luminescent composite during photoaging can be attributed to the accumulation of photoproducts in films. However, for PMMA/phosphor composites, the chemical structure and optical properties of such composites are kept for accelerated photoaging time up to 4000h.

M dwarfs constitute around 70% of all stars in the local Galaxy. Their multitude together with their long main-sequence lifetimes make them important for studies of global properties of the Galaxy such as the initial mass function or the structure and kinematics of stellar populations. In addition, the exoplanet community is showing an increasing interest for those small, cold stars. However, very few M dwarfs are well characterized, and in the case of exoplanetary systems the stellar parameters have a direct influence on the derived planet properties. Stellar parameters of M dwarfs are difficult to determine because of their low surface temperatures that result in an optical spectrum dominated by molecular lines. Most previous works have therefore relied on empirical calibrations. High-resolution spectrographs operating in the infrared, a wavelength region less affected by molecular lines, have recently opened up a new window for the investigation of M dwarfs. In the two first papers of this thesis we have shown that we can determine the metallicity, and in some cases the effective temperature, using synthetic spectral fitting with improved accuracy. This method is time consuming and therefore not practical or even feasible for studies of large samples of M dwarfs. When comparing our results from the high-resolution studies with available photometric calibrations we find systematic differences. In the third paper we therefore used our sample to determine a new photometric metallicity calibration. Compared to previous calibrations our new photometric calibration shows improved statistical characteristics, and our calibration gives similar results as spectroscopic calibrations. In a comparison with theoretical calculations we find a good agreement of the shapes and slopes of iso-metallicity lines with our empirical relation. Applying the photometric calibration to a sample of M dwarfs with confirmed exoplanets we find a possible giant planet-metallicity correlation for M dwarfs.

Cette thèse traite de l’utilisation de la couleur en vision par ordinateur. Deux problèmes sont étudiés : - l’estimation d’illuminants, - la mise en correspondance de descripteurs locaux pour la comparaison d’images couleur. Les surfaces achromatiques renvoient un spectre lumineux ayant la même distribution fréquentielle que le spectre de l’illuminant. Les détecter permet donc de recouvrer l'illuminant. En supposant que l’ensemble des couleurs que peut prendre un illuminant est limité (équation de Planck), il est possible de sélectionner les pixels appartenant à une surface potentiellement grise. Une méthode de vote est alors appliquée, permettant de sélectionner un ou plusieurs illuminants. L’algorithme final possède de nombreux avantages : il est efficace, intuitif, ne nécessite pas de phase d’apprentissage et requiert peu de paramètres, qui s’avèrent stables. De plus, la méthode de vote permet de s’adapter aux cas où plusieurs sources lumineuses éclairent la scène photographiée. Les descripteurs locaux sont des outils puissants pour comparer les images. Cependant, le rôle de la couleur dans l’étape d’appariement a fait l’objet de peu d’études. Le problème principal que nous considérons ici est celui de l’apport de la couleur pour l’appariement de descripteurs locaux. Un usage local de la couleur permet-il de désambiguïser les situations où la luminance seule est insuffisante ? Et si oui, dans quelles proportions ? Nous proposons quatre descripteurs permettant de décrire de manière détaillée le contexte local de points clés dans l’optique de les apparier. Les expériences proposées montrent clairement l’apport de la couleur pour la mise en correspondance locale. / This thesis addresses the use of color in image processing and computer vision. Two problems are studied: - illuminant estimation, - local descriptors matching for color images comparison. Achromatic surfaces are defined as surfaces reflecting a spectrum with the same frequency dsitribution than the illuminant. Consequently, revovering these surfaces allows to estimate the illuminant. Assuming that the range of colors taken by an illuminant is limited (Planck equation), it is possible to select the pixels belonging to a potentially gray surface. A voting procedure is then applied to select one or more illuminants. The proposed algorithm has many advantages: it is effective, intuitive, does not rely on a learning phase and requires only few parameters. In addition, the voting procedure can be adapted to handle cases where multiple light sources of different colors illuminate the scene. Local descriptors are powerful tools to compare images. However, few studies concern the influence of color in the matching step. The main problem that we consider here is the contribution of the color matching of local descriptors. Does the local use of color allow to disambiguate situations where the luminance alone is insufficient ? And if so, how much? We propose four descriptors to precisely describe the local context of key points in the matching step. The main idea developed here is that accurate information can only be obtained by describing both the color distributions and transitions between colors. The many experiences presented clearly show the positive contribution of color to the reliability of the local matching.

The accurate modeling of the variability of illumination in a class of images is a fundamental problem that occurs in many areas of computer vision and graphics. For instance, in computer vision there is the problem of facial recognition. Simply, one would hope to be able to identify a known face under any illumination. On the other hand, in graphics one could imagine a system that, given an image, the illumination model could be identified and then used to create new images. In this thesis we describe a method for learning the illumination model for a class of images. Once the model is learnt it is then used to render new images of the same class under the new illumination. Results are shown for both synthetic and real images. The key contribution of this work is that images of known objects can be re-illuminated using small patches of image data and relatively simple kernel regression models. Additionally, our approach does not require any knowledge of the geometry of the class of objects under consideration making it relatively straightforward to implement. As part of this work we will examine existing geometric and image-based re-lighting techniques; give a detailed description of our geometry-free face re-lighting process; present non-linear regression and basis selection with respect to image synthesis; discuss system limitations; and look at possible extensions and future work.