<p>У овој дисертацији разматране су објективне мере процене квалитета<br />слике и видеа са потпуним и делимичним референцирањем на изворни<br />сигнал. За потребе евалуације квалитета развијене су поуздане,<br />рачунски ефикасне мере, засноване на очувању информација о<br />градијенту. Мере су тестиране на великом броју тест слика и видео<br />секвенци, различитих типова и степена деградације. Поред јавно<br />доступних база слика и видео секвенци, за потребе истраживања<br />формиране су и нове базе видео секвенци са преко 300 релевантних<br />тест узорака. Поређењем доступних субјективних и објективних скорова<br />квалитета показано је да је објективна евалуација квалитета веома<br />сложен проблем, али га је могуће решити и доћи до високих<br />перформанси коришћењем предложених мера процене квалитета слике<br />и видеа.</p> / <p>U ovoj disertaciji razmatrane su objektivne mere procene kvaliteta<br />slike i videa sa potpunim i delimičnim referenciranjem na izvorni<br />signal. Za potrebe evaluacije kvaliteta razvijene su pouzdane,<br />računski efikasne mere, zasnovane na očuvanju informacija o<br />gradijentu. Mere su testirane na velikom broju test slika i video<br />sekvenci, različitih tipova i stepena degradacije. Pored javno<br />dostupnih baza slika i video sekvenci, za potrebe istraživanja<br />formirane su i nove baze video sekvenci sa preko 300 relevantnih<br />test uzoraka. Poređenjem dostupnih subjektivnih i objektivnih skorova<br />kvaliteta pokazano je da je objektivna evaluacija kvaliteta veoma<br />složen problem, ali ga je moguće rešiti i doći do visokih<br />performansi korišćenjem predloženih mera procene kvaliteta slike<br />i videa.</p> / <p>This thesis presents an investigation into objective image and video quality<br />assessment with full and reduced reference on original (source) signal. For<br />quality evaluation purposes, reliable, computational efficient, gradient-based<br />measures are developed. Proposed measures are tested on different image<br />and video datasets, with various types of distorsions and degradation levels.<br />Along with publicly available image and video quality datasets, new video<br />quality datasets are maded, with more than 300 relevant test samples.<br />Through comparison between available subjective and objective quality<br />scores it has been shown that objective quality evaluation is highly complex<br />problem, but it is possible to resolve it and acchieve high performance using<br />proposed quality measures.</p>

<p>Osnovni doprinos ove doktorske disertacije je razvoj algortima i sistema za objektivnu procenu vizuelnog kvaliteta slike uzimajući u obzir najvažnija moguća oštećenja kao što su zamućenje ivica (oštrina) i poremećaj prirodnog izgleda teksture objekata na slici sa jedne strane i uticaj sadržaja slike (procenta ivica u slici) na procenu kvaliteta sa druge strane. Dakle, hipoteza izneta u ovom radu je da je potreban multiparametarski pristup da bi se dobila objektivna procena kvaliteta slike koja je što približnija subjektivnoj proceni.</p>

<p>Основни допринос ове докторске дисертације је развој алгоритама за објективну процену визуелног квалитета слике синтетизоване применом ДИБР (Depth Image Based Rendering) техника које узрокују неуниформна изобличења у области ивица. Применом нелинеарних морфолошких филтара у мултирезолуционој декомпозицији слика код израчунавања предложене метрике, важне геометријске информације као што су ивице су добро очуване без помака и замућења у сликама на различитим скалама мултирезолуционе репрезентације. Израчунавањем МСЕ по подопсезима који садрже ивице, пиксел по пиксел, прецизно се мери разлика две мултирезолуционе репрезентације. Тако се највећи значај у процени квалитета додељује области ивица. Процене предложене метрике се добро поклапају са субјективним оценама.</p> / <p>Osnovni doprinos ove doktorske disertacije je razvoj algoritama za objektivnu procenu vizuelnog kvaliteta slike sintetizovane primenom DIBR (Depth Image Based Rendering) tehnika koje uzrokuju neuniformna izobličenja u oblasti ivica. Primenom nelinearnih morfoloških filtara u multirezolucionoj dekompoziciji slika kod izračunavanja predložene metrike, važne geometrijske informacije kao što su ivice su dobro očuvane bez pomaka i zamućenja u slikama na različitim skalama multirezolucione reprezentacije. Izračunavanjem MSE po podopsezima koji sadrže ivice, piksel po piksel, precizno se meri razlika dve multirezolucione reprezentacije. Tako se najveći značaj u proceni kvaliteta dodeljuje oblasti ivica. Procene predložene metrike se dobro poklapaju sa subjektivnim ocenama.</p> / <p>The main contribution of this doctoral thesis is the development of algorithms for objective<br />DIBR-synthesized view quality assessment. DIBR algorithms introduce nonuniform<br />geometric distortions affecting the edge coherency in the synthesized images.The non-linear<br />morphological filters used in multi-scale image decompositions of the proposed metric<br />maintain important geometric information such as edges across different resolution<br />levels.Calculating MSE pixel-by-pixel through subbands in which the edges are extracted,<br />the difference of the two multiresolution representations, the reference and the synthesized<br />image, is precisely measured. In that way the importance of edge areas which are prone to<br />synthesis artifacts is emphasized in the image quality assessment. The proposed metric has<br />very good agreement with human judgment.</p>

L'industrie automobile des systèmes embarqués critiques est confrontée de nos jours à une complexité croissante, tandis que les coûts, les performances en termes d'intelligence, les caractéristiques, les capacités et les délais de commercialisation de leurs produits sont constamment remises en question. Face à cela, l'objectif principal pour les constructeurs et fournisseurs automobiles devient désormais de contrôler la qualité et la fiabilité des systèmes mécatroniques et embarqués. L'existence de normes internationales comme le HIS Automotive SPICE et l’ISO26262 est une contrainte supplémentaire qu'ils doivent prendre en compte s’ils veulent atteindre cet objectif. De plus, assurer la bonne gestion de la sécurité et la qualité du produit ne suffit pas: il est essentiel de veiller à ce que nous produisons un système qui n'est pas seulement sécuritaire et bien, mais aussi que nous produisons le bon système. Cela induit donc une plus grande prise en compte des exigences.Dans cette thèse, nous traitons le challenge du développement des systèmes embarqués automobiles suivant l’Ingénierie Dirigée par les Modèles (IDM) qui répondent aux exigences des utilisateurs et des standards du domaine et qui permettent de maîtriser davantage la qualité des produits développés. Le problème à résoudre a été abordé sur plusieurs phases qui sont ensuite utilisés conjointement. En premier, nous définissons un métamodèle fusionnant les approches orienté qualité produit et qualité processus selon respectivement les normes ISO26262 et SPICE. Puis, dans un but de certification, nous proposons une méthodologie générique basée sur ce métamodèle commun où une évaluation du processus de développement induite par l’HIS standard ainsi qu’une évaluation de la sécurité fonctionnelle induite par l’ISO2626 sont simultanément effectuées. Ce résultat est traduit au travers de la définition d'un framework outillé où nous appliquons la méthode d'évaluation propre au standard SPICE. En deuxième phase, nous définissons un métamodèle pour gérer les actifs de sécurité concernant ces normes automobiles au niveau produit. Ce métamodèle définit comment capturer les exigences et l’architecture d’un système de telle manière qu'ils puissent être traçables entre eux et également traçables depuis des documents de spécifications d’origine. Enfin, une approche à base de modèle où l'interaction des modèles de processus et le produit est géré afin de répondre aux besoins identifiés dans la première phase est développé pour soutenir la gestion de projet. L'approche utilise la modélisation des processus et leur mesure pour améliorer le contrôle et le suivi de projet et de réduire par la même les coûts et les fréquences de replanification.Les avantages de la contribution sont démontrés sur une application pilote automobile, validant ainsi le travail de recherche au vue des faiblesses identifiées préalablement dans le contexte. / The embedded safety-critical systems industry is facing an exponential increase in the complexity and variety of systems and devices while costs, performance in terms of intelligence, features, capacities and time to market are constantly challenged. The main objective for automotive manufacturers and suppliers is now becoming the control of quality and the dependability of embedded and mechatronic systems. The existence of internationally recognized standards such as the Automotive SPICE and ISO26262 is a further constraint that must be managed to meet this objective. Nevertheless, ensuring sound management of safety and viewpoints is insufficient. It is also essential to ensure that we produce a system that is not only compliant and well-defined, but also that we produce the “right” system. Therefore, this leads to greater consideration of the requirements.In this thesis, we address the challenge of development of automotive embedded systems following the model-driven engineering paradigm that meet the user needs and the regulatory constraints of the domain and that further mastered the quality of developed product. We resolve the problem in many steps which are subsequently used jointly. In the first phase, we define a merging approach which embodies a product quality and process quality approaches regarding the ISO26262 and SPICE standards following the model-driven engineering paradigm. Then, in a certification assessment purpose, we propose a generic methodology where an HIS assessment and a functional safety audit is simultaneously performed without altering their original meanings. This commitment results into the definition of a tooled framework where we apply the SPICE assessment method to the common metamodel defined from the merging work. In a second phase, we define a metamodel for manage safety assets regarding these automotive standards at product level. This metamodel defines how the requirements and architecture of a system can be captured in such a way that they can be traceable from each other and from origin specifications documents. Finally, a model-based approach where the interaction of process and product models is managed to address requirements identified in the precedent phases is developed to support project management. The approach uses process modeling and measurement to improve the control and the monitoring of project and to reduce the cost and frequency of re-planning.The benefits of the contribution are demonstrated on an ongoing automotive pilot application, thereby validating the research work against the weaknesses identified prealably in the context.

Speech as a medium for communication has always been important in its ability to convey our ideas, personality and emotions. It is therefore not strange that Quality of Experience (QoE) becomes central to any business relying on voice communication. Using Unified Communication (UC) systems, users can communicate with each other in several ways using many different devices, making QoE an important aspect for such systems. For this thesis, automatic methods for assessing speech quality of the voice calls in Briteback’s UC application is studied, including a comparison of the researched methods. Three methods all using a Gaussian Mixture Model (GMM) as a regressor, paired with extraction of Human Factor Cepstral Coefficients (HFCC), Gammatone Frequency Cepstral Coefficients (GFCC) and Modified Mel Frequency Cepstrum Coefficients (MMFCC) features respectively is studied. The method based on HFCC feature extraction shows better performance in general compared to the two other methods, but all methods show comparatively low performance compared to literature. This most likely stems from implementation errors, showing the difference between theory and practice in the literature, together with the lack of reference implementations. Further work with practical aspects in mind, such as reference implementations or verification tools can make the field more popular and increase its use in the real world.

Text in English / Undergraduate students at the Eduardo Mondlane University (UEM) experience great difficulty in developing their translation competence during their training period. On the one hand, they show many signs of poor quality during their training when they accomplish practical translation assignments on and off-campus. On the other hand, the quality of the final work submitted in partial fulfilment of the requirements for the Bachelor Honours Degree (Licenciatura) suggests that most students fail to go beyond the minimum standards with regard to translation competence. Yet, comprehensive research aimed at understanding factors hindering translation competence at the UEM has been scanty. Besides, the few studies available have some significant lacunae, for they focus on Portuguese language development; are limited to error analysis, and often ignore students‟ perspectives. Consequently, the problem of finding a balanced approach to developing students‟ translation competence has remained unaddressed. Therefore, this study has been conducted with the purpose of investigating the translation competence of the UEM undergraduates in order to establish the major obstacles to their translation competence development, and consequently come up with suggestions for improving the current translator-training degree programme. The study design uses qualitative methods translated into a case study approach. First, questionnaire data is analysed to gain lecturers‟ and students‟ perspectives on translation competence at the UEM. Second, a students‟ error typology is developed following a macro- and micro-textual analysis of their translations. The study concludes that substandard bilingual skills, compounded by curricular, attitudinal/motivational, pedagogical and infrastructural/instrumental factors, negatively impact the students‟ development of translation competence at the UEM and that addressing these may pave the way towards improving the current translation-training programme. / Linguistics and Modern Languages / M. A. (Linguistics)

Diffusion Tensor Imaging (DTI) is a magnetic resonance imaging (MRI)-based technique that has frequently been used for the identification of brain biomarkers of neurodevelopmental and neurodegenerative disorders because of its ability to assess the structural organization of brain tissue. In this work, I present (1) preclinical findings of a longitudinal DTI study that investigated asymptomatic high school football athletes who experienced repetitive head impact and (2) an automated pipeline for assessing the quality of DTI images that uses a convolutional neural network (CNN) and transfer learning. The first section addresses the effects of repetitive subconcussive head trauma on the white matter of adolescent brains. Significant concerns exist regarding sub-concussive injury in football since many studies have reported that repetitive blows to the head may change the microstructure of white matter. This is more problematic in youth-aged athletes whose white matter is still developing. Using DTI and head impact monitoring sensors, regions of significantly altered white matter were identified and within-season effects of impact exposure were characterized by identifying the volume of regions showing significant changes for each individual. The second section presents a novel pipeline for DTI quality control (QC). The complex nature and long acquisition time associated with DTI make it susceptible to artifacts that often result in inferior diagnostic image quality. We propose an automated QC algorithm based on a deep convolutional neural network (DCNN). Adaptation of transfer learning makes it possible to train a DCNN with a relatively small dataset in a short time. The QA algorithm detects not only motion- or gradient-related artifacts, but also various erroneous acquisitions, including images with regional signal loss or those that have been incorrectly imaged or reconstructed.

<a>Global trends and factors, such as the increased level of globalization, climate change, resource scarcity, and awareness of social and environmental responsibilities, as well as fiercer competition and lower profit margins in all industries, force organizations to act to retain, regain, or sustain their competitive advantages for long-term survival. These trends and factors are historically known to bring about innovations that drive the evolution of industries. Sustainability is considered to be such an innovation to achieve fiscally sound, environmentally conscious, and socially progressive organizations and supply chains. Sustainable Development and Sustainability notions are among trending topics of 21st century. Elevated sustainability concerns of various stakeholders have been forcing members of all industries to evolve into their more environmentally and socially responsible versions. This study was initiated through a comprehensive literature review phase that reviewed 477 articles published in five major databases from 1990 to 2018. The purpose of this review was to assess the current state-of-the art on the subject of lean-driven sustainability. Based on descriptive and contextual analysis, synergies, divergences, and the extent of two-way permeability of lean and sustainability concepts from the perspective of intra- and inter-organizational operations were identified along with future research opportunities. Fundamental strengths and weaknesses of both concepts, existing strong synergies and untapped potential, along with their key contributors, the potential-use cases of lean tools to derive sustainable solutions are highlighted in this review. Next, based on the findings of systematic literature review, an innovative, holistic, versatile and scalable tool was developed to assess and benchmark sustainability performance of organizations and supply chains. The proposed framework was established upon trivet structure of Triple Bottom Line philosophy and fueled by Lean, Six-Sigma and Life Cycle Assessment (LCA) methodologies for accurate and effective measurement of sustainability performance. Completeness of the framework was ensured through development of first-generation Key Performance Indicator (KPI) pool with 33 indicators, a unique work environment assessment mechanism for safety and environmental protection issues in terms of 11 risk categories and by construction of an ownership structure for ease of framework deployment. Proposed framework is expected to help with true sustainability performance improvement and benchmarking objectives at a range of business levels from facility to sectoral operations. Upon completion of the development phase, the Sustainability Benchmarking Tool (SBT) Framework was validated at the facility level within the context of value-added wood products manufacturing. Strengths and weaknesses of the system were identified within the scope of Bronze Frontier maturity level of the framework and tackled through a six-step analytical and quantitative reasoning methodology. The secondary objective of the validation phase was to document how value-added wood products industries can take advantage of natural properties of wood to become frontiers of sustainability innovation. In the end, True Sustainability performance of the target facility was improved by 2.37 base points, while economic and environmental performance was increased from being a system weakness to achieving an acceptable index score benchmark of 8.41 and system strength level of 9.31, respectively. Social sustainability score increased by 2.02 base points as a function of better gender bias ratio. The financial performance of the system improved from a 33% loss to 46.23% profit in the post-improvement state. Reductions in CO₂ emissions (55.16%), energy consumption (50.31%), solid waste generation (72.03%), non-value-added-time (89.30%) and cost performance (64.77%) were other significant achievements of the study. In the end, SBT Framework was successfully validated at the facility level and target facility evolved into its leaner, cleaner and more responsible version of itself. Furthermore, manufacturing industries of all sorts are key stakeholders, which rely on universities to satisfy the demand for competent workforce. Society also expects universities to educate youth and contribute to their self-development by achieving both, scientific and intellectual knowledge saturation. To expand the contribution of the study to the body of knowledge in the fields of Sustainability and Modern Management techniques, an undergraduate level course curriculum that integrates modern management techniques and sustainability concepts with wood products industry dynamics was developed. Students’ pre- and post-education awareness of, and familiarity with sustainability, potential consequences of ignored sustainability issues, modern management techniques, global trends, innovation waves, and industry evolution were compared through a seventeen-question survey. Results showed that course content was successful at increasing sustainability awareness at both overall and individual sustainability pillar levels, At the end, 100% of students were able to develop complete understanding of various modern management techniques and stated that they felt confident to apply learnt skills to real life issues within their profession upon graduation. Overall, this study empirically documented how synergies between Lean, Sustainability, Six-Sigma and Life Cycle Assessment concepts outweigh their divergences, demonstrated viability of SBT Framework and presented a proven example of modern management techniques powered transdisciplinary sustainability curriculum.</a>

Although Eduardo Mondlane University (UEM) has the longest history of BA Honours level translator and interpreter training in Mozambique, the university still lacks an effective model for the development of translation and interpreting competence in students. To address this problem, the present study seeks to find a practical model for the training of Mozambican professional translators and interpreters at BA Honours level that can guide the design of conducive curricula. The critical question the study attempts to answer is: What model for developing translation and interpreting competence could lead to an effective curriculum design that best meets the employment needs of Mozambican students? To this end, the study has been designed as action-research because this design enables better understanding and improvement of training processes (Cravo & Neves 2007). Three data collection tools are used to generate both qualitative and quantitative data from over 120 participants, namely: (i) a survey, (ii) an English translation test and (iii) a sample of archived Portuguese translations produced by former students. The survey findings suggest the need for a model whereby translators and interpreters are trained simultaneously within the same programme. Moreover, the results of macro- and micro-textual analysis show that, overall, the translation competence of former students is poor, suggesting that the current curriculum at UEM is failing to produce BA Honours translation/interpreting professionals. The proposed solution would be a curriculum based on a new integrated translation and interpreting competence development model with the following four pillars: communicative competence, general knowledge, strategic competence and service provision. / Nangona iyunivesithi iEduardo Mondlane University (UEM) inembali kwizifundo zesidanga esiphakamileyo (BA Honours) kuqeqesho lwabaguquleli neetoliki eMozambique, le yunivesithi ayikabi namzekelo unguwo nosebenzayo ekuphuhliseni izakhono zokuguqula nokutolika kubafundi bayo. Ukukhawulelana nale ngxaki olu phando lujolise ekufumaneni owona mzekelo usebenzayo onokusetyenziswa nonokuthi ube sisikhokelo kuyilo lwekharityhulam yesidanga esiphakamileyo iBA Honours esithatha iminyaka emine ekuqeqesheni abafundi babe ngabaguquleli neetoliki eziphume izandla zaseMozambique. Umbuzo ongundoqo nozanywa ukuphendulwa lolu phando ngulo: Ngowuphi umzekelo wokuphuhlisa izakhono zokuguqulela nokutolika onokukhokelela kuyilo lwekharityhulam esebenzayo nefezekisa iimfuno zengqesho zabafundi baseMozambique? Kungoko olu phando luye lwasebenzisa indlela yokuphanda ekuthiwa yi-action research eyenza kube lula ukuqonda nokuphucula iinkqubo zoqeqesho (Cravo & Neves 2007). Kuye kwasetyenziswa iindlela ezintathu zophando ukufumana ulwazi kubathathi-nxaxheba abali-120 ezizezi: (1) uvavanyo lwezimvo, (ii) uvavanyo lwenguqulelo yesicatshulwa esibhalwe ngesiNgesi kunye (iii) neenguqulelo zesiPhuthukezi ezenziwe ngabafundi bangaphambili ezigciniweyo. Iziphumo zeemvavanyo zezimvo zibonisa ukuba kufuneka kukho umzekelo onokusetyenziswa ekuqeqesheni abaguquleli neetoliki ngaxeshanye phantsi kwenkqubo enye. Ukongeza koku, iziphumo zohlalutyo lwesicatshulwa zibonise ukuba izakhono zokuguqulela zabafundi bangaphambili azikho mgangathweni, nto leyo ethetha ukuba ikharityhulam esetyenziswayo eUEM iyasilela ekukhupheni abaguquleli neetoliki ezinobuchule neziziincutshe. Isisombululo esicetyiswayo ke ngoko, sesokuba kubekho ikharityhulam ehlangeneyo eza kuphuhlisa izakhono zabaguquleli neetoliki esekelwe kwiintsika ezine ezizezi: isakhono sonxibelelwano, ulwazi jikelele, isakhono sobuchule kunye nokunikezwa kweenkonzo. / Nakuba iNyuvesi i-Eduardo Mondlane University (i-UEM) isinomlando omude kakhulu wokuqeqesha abahumushi notolika ezingeni leziqu ze-BA Honours eMozambique, le nyuvesi ayikabi nayo indlela esebenzayo yokuthuthukisa amakhono ezitshudeni kwezokuhumusha nokutolika. Ukubhekana nale nkinga lolu cwaningo kuhloswe ngalo ukuthola indlela esebenzayo yokuqeqesha ongoti babahumushi notolika baseMozambique ezingeni leziqu ze-BA Honours (iziqu zeminyaka emine) ezingahlahla indlela yokuklama uhlelo lwezifundo olungasiza kule nkinga. Umbuzo omkhulu lolu cwaningo oluzama ukuwuphendula ngothi: Iyiphi indlela yokuthuthukisa amakhono okuhumusha nokutolika engaholela ekuklanyweni kohlelo lwezifundo olungahlangabezana nezidingo zokuqasheka kwezitshudeni zaseMozambique? Ukufeza le nhloso, lolu cwaningo lusebenzisa uhlelo lokucwaninga olubizwa nge-action-research ngoba luyasiza ekuqondeni kangcono nasekuthuthukiseni inqubo yokuqeqesha (Cravo & Neves 2007). Kusetshenziswa amathuluzi amathathu okuqoqa imininingobunjalo nemininingobuningi evela kubabambiqhaza abangaphezu kwabayi-120, okuyilokhu: (i) ngohlolocwaningo (i-survey), (ii) ngesivivinyo sesihumusho sesiNgisi (iii) nangesampula lezihumusho zesiPutukezi zabafundi baphambilini. Okutholakale kulo uhlolocwaningo kuveza isidingo sendlela yokuqeqesha abahumushi notolika kanyekanye ohlelweni lokufundiswa olufanayo. Ngaphezu kwalokho, imiphumela yokuhlaziywa kwemibhalo nezimo eyabhalwa ngaphansi kwazo ibonisa ukuthi, ngokubanzi, amakhono okuhumusha abafundi baphambilini awamahle, okuyinkomba yokuthi uhlelo lwezifundo lwamanje e-UEM luyehluleka ukukhiqiza ongoti bokuhumusha nokutolika ezingeni le-BA Honours. Isixazululo esiphakanyiswayo ngesendlela entsha esuselwa ekuthuthukiseni amakhono edidiyela ukuhumusha nokutolika, enalezi zinsika ezine ezilandelayo: amakhono kwezokuxhumana, ulwazi ngokubanzi, ikhono lokusebenzisa amasu, nokuhlinzekwa kwezidingo. / Linguistics and Modern Languages / D. Phil. (Languages, Linguistics and Literature)

No description available.