Cette thèse porte sur l’étude des stratégies d’amélioration de la qualité d’image dans les systèmes de communication sans fil et sur la conception de nouvelles métriques d’évaluation de la qualité. Tout d'abord, une nouvelle métrique de qualité d'image à référence réduite, basée sur un modèle statistique dans le domaine des ondelettes complexes, a été proposée. Les informations d’amplitude et de phase relatives des coefficients issues de la transformée en ondelettes complexes sont modélisées à l'aide de fonctions de densité de probabilité. Les paramètres associés à ces fonctions constituent la référence réduite qui sera transmise au récepteur. Ensuite, une approche basée sur les réseaux de neurones à régression généralisée est exploitée pour construire la relation de cartographie entre les caractéristiques de la référence réduite et le score objectif.Deuxièmement, avec la nouvelle métrique, une nouvelle stratégie de décodage est proposée pour la transmission d’image sur un canal de transmission sans fil réaliste. Ainsi, la qualité d’expérience (QoE) est améliorée tout en garantissant une bonne qualité de service (QoS). Pour cela, une nouvelle base d’images a été construite et des tests d’évaluation subjective de la qualité de ces images ont été effectués pour collecter les préférences visuelles des personnes lorsqu’elles sélectionnent les images avec différentes configurations de décodage. Un classificateur basé sur les algorithmes SVM et des k plus proches voisins sont utilisés pour la sélection automatique de la meilleure configuration de décodage.Enfin, une amélioration de la métrique a été proposée permettant de mieux prendre en compte les spécificités de la distorsion et la préférence des utilisateurs. Pour cela, nous avons combiné les caractéristiques globales et locales de l’image conduisant ainsi à une amélioration de la stratégie de décodage.Les résultats expérimentaux valident l'efficacité des métriques de qualité d'image et des stratégies de transmission d’images proposées. / This thesis focuses on the study of image quality strategies in wireless communication systems and the design of new quality evaluation metrics:Firstly, a new reduced-reference image quality metric, based on statistical model in complex wavelet domain, has been proposed. The magnitude and the relative phase information of the Dual-tree Complex Wavelet Transform coefficients are modelled by using probability density function and the parameters served as reduced-reference features which will be transmitted to the receiver. Then, a Generalized Regression Neural Network approach is exploited to construct the mapping relation between reduced-reference feature and the objective score.Secondly, with the new metric, a new decoding strategy is proposed for a realistic wireless transmission system, which can improve the quality of experience (QoE) while ensuring the quality of service (QoS). For this, a new database including large physiological vision tests has been constructed to collect the visual preference of people when they are selecting the images with different decoding configurations, and a classifier based on support vector machine or K-nearest neighboring is utilized to automatically select the decoding configuration.Finally, according to specific property of the distortion and people's preference, an improved metric has been proposed. It is the combination of global feature and local feature and has been demonstrated having a good performance in optimization of the decoding strategy.The experimental results validate the effectiveness of the proposed image quality metrics and the quality strategies.

Cette thèse concerne la restauration aveugle d’images (formulée comme un problème inverse mal-posé et mal-conditionné), en considérant particulièrement les systèmes SIMO. Dans un premier temps une technique d’identification aveugle de ce système où l’ordre du canal est inconnu (surestimé) est introduite. Nous introduisons d’abord une version simplifiée à coût réduit SCR de la méthode des relations croisées (CR). Ensuite, une version robuste R-SCR basée sur la recherche d’une solution parcimonieuse minimisant la fonction de coût CR est proposée. La restauration d’image est ensuite assurée par une nouvelle approche inspirée des techniques de décodage des signaux 1D et étendue ici aux cas de la restauration d’images en se basant sur une recherche arborescente efficace (algorithme ‘Stack’). Plusieurs améliorations de la méthode ‘Stack’ ont été introduites afin de réduire sa complexité et d’améliorer la qualité de restauration lorsque les images sont fortement bruitées. Ceci en utilisant une technique de régularisation et une approche d’optimisation all-at-once basée sur la descente du gradient qui permet de raffiner l’image estimée et mieux converger vers la solution optimale. Ensuite, les mesures de la qualité d’images sont utilisées comme fonctions de coûts (intégrées dans le critère global) et ce afin d’étudier leur potentiel pour améliorer les performances de restauration. Dans le contexte où l’image d’intérêt est corrompue par d’autres images interférentes, sa restauration nécessite le recours aux techniques de séparation aveugle de sources. Pour cela, une étude comparative de certaines techniques de séparation basées sur la propriété de décorrélation au second ordre et la parcimonie est réalisée. / This thesis concerns the blind restoration of images (formulated as an ill-posed and illconditioned inverse problem), considering a SIMO system. Thus, a blind system identification technique in which the order of the channel is unknown (overestimated) is introduced. Firstly, a simplified version at reduced cost SCR of the cross relation (CR) method is introduced. Secondly, a robust version R-SCR based on the search for a sparse solution minimizing the CR cost function is proposed. Image restoration is then achieved by a new approach (inspired from 1D signal decoding techniques and extended here to the case of 2D images) based on an efficient tree search (Stack algorithm). Several improvements to the ‘Stack’ method have been introduced in order to reduce its complexity and to improve the restoration quality when the images are noisy. This is done using a regularization technique and an all-at-once optimization approach based on the gradient descent which refines the estimated image and improves the algorithm’s convergence towards the optimal solution. Then, image quality measurements are used as cost functions (integrated in the global criterion), in order to study their potential for improving restoration performance. In the context where the image of interest is corrupted by other interfering images, its restoration requires the use of blind sources separation techniques. In this sense, a comparative study of some separation techniques based on the property of second-order decorrelation and sparsity is performed.

The objective of this research is to develop and design front-end analog circuits for Capacitive Micromachined Ultrasound Transducers (CMUTs) and optical grating MEMS microphone. This work is motivated by the fact that with micro-scaling, MEMS sense capacitance gets smaller in a CMUT array element for intravascular ultrasound imaging, which has dimensions of 70um x 70um and sub pico-farad capacitance. Smaller sensors lead to a lower active-to-parasitic ratio and thus, degrads sensitivity. Area and power requirements are also very stringent, such as the case of intravascular catheter implementations with CMOS-First CMUT fabrication approach. In this implementation, capacitive feedback charge amplifier is an alternative approach to resistive feedback amplifiers. Capacitive feedback charge amplifier provides high sensitivity, small area, low distortion and saving power. This approach of charge amplifiers is also suitable in capacitive microphones where it provides low power and high sensitivity. Another approach to overcome capacitive detection challenges is to implement optical detection. In the case of biomimetic microphone structure, optical detection overcomes capacitive detection's thermal noise issues. Also with micro-scaling, optical detection overcomes the increased parasitics without any sensitivity degradation, unlike capacitive detection. For hearing aids, along with sensitivity, battery life is another challenge. We propose the use of 1-bit front-end sigma-delta ADC for overall improved hearing aid power efficiency. Front-end interface based on envelope detection and synchronous detection schemes have also been designed. These interface circuits consume currents in microampere range from a 1.5V battery. Circuit techniques are used for maximizing linear range and signal handling with low supplies. The entire front end signal processing with Vertical Cavity Surface Emitting Laser (VCSEL) drivers, photodiodes, filters and detectors is implemented on a single chip in 0.35um CMOS process.

During hospitalisation in the SCBU premature neonates may undergo a significant number of radiographic procedures to assist mainly in the diagnosis and management of lung diseases, which represent one of the most life threatening conditions in the newborn. Additionally, repeated radiographs are required to confirm correct positioning of tubes and catheters inserted during the course of their management. Special attention should be paid in optimising exposures, as neonates have an increased risk of radiation induced malignancy compared to adults due to the high radiosensitivity of mitotic cells and their longer life expectancy. Another consideration in the frequent imaging of neonates, while in the SCBU, is that handling of the neonate should be minimised. Current clinical practice mainly involves positioning the cassette on the bed directly behind the neonate. However, lifting or moving the infant to position the cassette, can lead to hypoxia, bradycardia, cerebral haemorrhage and could cause accidental dislodgment of tubes, lines and probes. Additionally, medical and nursing issues include increased risk of cross infection and changes in the stable microenvironment of the incubator, e.g. humidity and temperature. Modern incubators incorporate an imaging tray under the bed to facilitate placement of the radiographic cassette without the need to disturb the infant. Paediatricians and nurses urge the use of the tray for the benefit of the neonate. One the other hand, great concerns form the radiographers were expressed, regarding increased dose and poor image quality due to greater attenuation of the beam, as well as for repeat exposures due to artifacts or misalignment. The current study was divided in three parts. The first objective of the study was to perform a survey of the doses encountered in two Special Care Baby Units, one at the University Hospital of Patras in Greece and the other one at the Royal Victoria Infirmary in Newcastle upon Tyne in the UK. The first part of this study was performed in Greece and an optimized radiographic protocol for film-screen was suggested according to neonatal birth weight. In the second part, realized in the UK, the use of the incubator imaging tray is investigated for positioning the CR cassette, with regards to detector dose, exposure index, image quality and radiation dose to patient. Part three is studying the effect of the new tissue radiosensitivity factors, published in the new report from the International Commission on Radiological Protection 2007, on the effective dose conversion coefficients in computed tomography examinations in paediatric patients. / Τα νεογνά, και ιδιαίτερα τα πρόωρα (διάρκεια κύησης έως και 24 εβδομάδες και βάρος γέννησης έως και 600 g) αμέσως μετά την γέννηση τους διανύουν μια περίοδο υψηλού κινδύνου και συχνά έχουν να αντιμετωπίσουν μια ποικιλία από σοβαρές έως και απειλητικές για τη ζωή επιπλοκές στην υγεία τους, που συχνά απαιτούν ειδική φροντίδα κατά τη διάρκεια των πρώτων ημερών στην Μονάδα Εντατικής Θεραπείας Νεογνών. Τόσο η καθυστερημένη ενδομήτρια ανάπτυξη, όσο και ο πρόωρος τοκετός αποτελούν σημαντικές αιτίες για την εμφάνιση αναπνευστικών ή καρδιαγγειακών προβλημάτων. Το ποσοστό επιβίωσης αυτών των νεογνών, ωστόσο, έχει αυξηθεί δραματικά κατά τις τελευταίες δύο δεκαετίες, που βασίζεται τόσο στην άµεση διάγνωση όσο και στην έγκαιρη και αποτελεσματική θεραπεία. Η ακτινογράφηση νεογνών είναι ένα απολύτως αναγκαίο και ισχυρό «εργαλείο» που συμβάλλει σηµαντικά όχι µόνο στην έγκαιρη αρχική διάγνωση και εκτίµηση της ασθένειας (π.χ αναπνευστική δυσχέρεια, εισρόφηση µηκωνίου, βρογχοπνευµονική δυσπλασία, πνευµοθώρακας, νεκρωτική εντεροκολίτιδα κτλ) αλλά και στην τοποθέτηση και επιβεβαίωση της σωστής θέσης των διαδερµικών ενδοφλέβιων κεντρικών γραµµών, οµφαλικών καθετήρων και ενδοτραχειακών σωλήνων που απαιτούνται κατά την αντιµετώπιση των συµπτωµάτων αυτών καθώς και κατά την παρακολούθηση της θεραπείας. Συνεπώς, ανάλογα µε τα κλινικά συμπτώματα που παρουσιάζουν, τα πρόωρα νεογνά υποβάλλονται σε ένα αρκετά σηµαντικό αριθµό ακτινογραφικών εξετάσεων θώρακος και κοιλίας (έως και > 60 ακτινογραφίες) κατά τις πρώτες µέρες ζωής. Επιπλέον, τα νεογνά πιθανόν να υποβληθούν σε μια ποικιλία από άλλες ακτινολογικές τεχνικές κατά τη διαχείριση της κατάστασής τους. Αν και η αξονική τομογραφία σε νεογνά (computed tomography- CT) π.χ. για την εκτίμηση των συγγενών ανωμαλιών της καρδιάς ή τραύμα στο κεφάλι, είναι σχετικά σπάνια σε σύγκριση με την ακτινογράφηση, ενόψει των υψηλών δόσεων που συνδέονται με την τεχνική αυτή, είναι σημαντικό για τους ακτινολόγους και τους τεχνολόγους να είναι σε θέση να εκτιμήσουν την δόση και τα στοχαστικά αποτελέσματα, όπως την πιθανότητα εμφάνισης καρκίνου η λευχαιμίας, εξαιτίας της ακτινοβολίας. Κατά τη διάρκεια εξετάσεων CT, η ενέργεια εναποτίθεται στα διάφορα όργανα με ένα περίπλοκο τρόπο, ανάλογα με την ακτινοβολούμενη ανατομική περιοχή και τις διαστάσεις του σώματος του ασθενή. Η εκτίμηση του κινδύνου καρκίνου απαιτεί γνώση των εν λόγω δόσεων στα όργανα. Ως εκ τούτου, κρίθηκε απαραίτητο να διερευνηθεί και να επαναπροσδιοριστεί μια μεθοδολογία δοσιμέτρησης, με σκοπό να συμβάλλει στη βελτιστοποίηση των ακτινολογικών τεχνικών στις αξονικές εξετάσεις νεογνών. Η ακτινοευαισθησία κάθε ιστού είναι ευθέως ανάλογη του ρυθµού εξάπλωσης και πολλαπλασιασµού των κυττάρων. Τα νεογνά είναι έως και δέκα φορές πιο ευαίσθητα σε σχέση µε τους ενήλικες στις χρωµοσωµατικές καταστροφικές συνέπειες της ακτινοβολίας, εξ’ αιτίας της υψηλής µιτωτικής δραστηριότητας των κυττάρων τους. Επιπλέον, κρίσιµα και ακτινοευαίσθητα όργανα, όπως ο µαστός, ο θυρεοειδής αδένας, οι γονάδες και ένα µεγάλο τµήµα του αιµοπαραγωγικού µυελού των οστών βρίσκονται κοντά ή εντός της πρωτογενούς δέσµης και ακτινοβολούνται απ’ ευθείας. Επίσης, λόγω του µεγαλύτερου προσδόκιµου ζωής των νεογνών σε σχέση µε οποιαδήποτε άλλη οµάδα ασθενών, υπάρχει μεγαλύτερη περίοδος της πιθανής εµφάνισης κακοήθειας εξ’ αιτίας της ακτινοβολίας. Ένα άλλο στοιχείο σχετικό με την συχνή απεικόνιση των νεογνών, κατά την παραμονή τους στην μονάδα, είναι ότι ο χειρισμός και η αλληλεπίδραση με το νεογνό θα πρέπει να ελαχιστοποιούνται. Τα πρόωρα νεογνά δέχονται ιατρική φροντίδα σε θερμοκοιτίδες, οι οποίες είναι σχεδιασμένες για να παρέχουν το βέλτιστο ελεγχόμενο μικροπεριβάλλον, ιδανικό για το μωρό, όπου οι ζωτικές λειτουργίες του μπορούν να παρακολουθούνται προσεκτικά συνεχώς. Κατά την ακτινογράφηση, στην τρέχουσα κλινική πρακτική, η κασέτα τοποθετείται κυρίως πάνω στο κρεβάτι, ακριβώς πίσω από το νεογνό. Ωστόσο, η ανάγκη για μετακίνηση και ανύψωση του νεογνού μπορεί να οδηγήσει σε υποθερμία, υποξία, βραδυκαρδία, εγκεφαλική αιμορραγία, καθώς και μεταφορά μολύνσεων και τυχαία εκτόπιση των κεντρικών γραμμών και καθετήρων. Οι σύγχρονες θερμοκοιτίδες είναι εφοδιασμένες με ένα μια υποδοχή/συρόμενο δίσκο απεικόνισης (imaging tray) κάτω από το κρεβάτι που διευκολύνει την τοποθέτηση της ακτινολογικής κασέτας με σκοπό την ελάχιστη ενόχληση του μωρού και μεταβολή των συνθηκών μέσα στην θερμοκοιτίδα. Για ακτινογραφίες που λαμβάνονται χρησιμοποιώντας το συρόμενο δίσκο απεικόνισης, λόγω της εξασθένησης που προκαλείται από το κρεβάτι, το στρώμα και τα άλλα υλικά που παρεμβαίνουν στην δέσμη, πιθανόν να απαιτείται αύξηση των παραμέτρων έκθεσης, με αποτέλεσμα το νεογνό να εκτεθεί σε υψηλότερη δόση. Το κρεβάτι συνήθως δεν είναι κατασκευασμένο από υλικά χαμηλής απορρόφησης, αλλά από ένα απλό πλαστικό υλικό, έτσι ώστε υπάρχουν ισχυρά επιχειρήματα που να υποστηρίζουν ότι η τοποθέτηση της κασέτας πίσω από το νεογνό είναι πιο συνεπής με την αρχή της βελτιστοποίησης. Η παρούσα μελέτη αποτελείται από τρία βασικά μέρη. Ο πρώτος στόχος της μελέτης ήταν να πραγματοποιηθεί μια έρευνα σχετικά με τον καθορισμό της δόσεων που σχετίζονται με ακτινογραφικές εξετάσεις θώρακος στην Μονάδα Εντατικής Θεραπείας Νεογνών στο Πανεπιστημιακό Νοσοκομείο της Πάτρας στην Ελλάδα, η σύγκριση των δόσεων αυτών με τα διεθνή διαγνωστικά επίπεδα αναφοράς και ανάπτυξη και η παρουσίαση ενός βελτιστοποιημένου ακτινογραφικού πρωτοκόλλου με χρήση ακτινολογικού φιλμ, ανάλογα με το σωματικό βάρος γέννησης του νεογνού, βασισμένη σε κλινικές εικόνες νεογνών. Δεδομένου ότι η ψηφιακή ακτινογράφηση είναι σήμερα η πλέον χρησιμοποιούμενη μορφή απεικόνισης σε πολλά κέντρα, η μελέτη αναπτύχθηκε περαιτέρω ώστε να συγκρίνει τις τεχνικές και τις παραμέτρους έκθεσης που εφαρμόζονται κλινικά στην μονάδα εντατικής θεραπείας νεογνών στο Royal Victoria Infirmary (RVI) στο Newcastle Upon Tyne, στο Ηνωμένο Βασίλειο. Επιπλέον, ερευνήθηκε η επίδραση της χρήσης του συρόμενου δίσκου απεικόνισης της θερμοκοιτίδας για την τοποθέτηση της κασέτας κατά τη ακτινογράφηση θώρακος, σε σχέση με τη δόση στη κασέτα, το δείκτη έκθεσης της ψηφιακής εικόνας, την ποιότητα της εικόνας και των επιπτώσεων από τη χρήση του δίσκου στο νεογνό. Η μελέτη πραγματοποιήθηκε βασισμένη σε εικόνες από ανθρωπόμορφο ομοίωμα νεογνού με ψηφιακο ανιχνευτή (computed radiography-CR). Στο τρίτο μέρος μελετάται και επαναπροσδιορίζεται μια μεθοδολογία δοσιμέτρησης σε εξετάσεις CT, καθώς και η επίδραση των νέων παραγόντων ακτινοευαιθησίας των ιστών, όπως δημοσιεύθηκαν στην νέα έκθεση από τη Διεθνή Επιτροπή Ακτινοπροστασίας το 2007 (International Radiological Protection Board-ICRP), στην ενεργό δόση (effective dose- E) και στους συντελεστές μετατροπής της δόσης (effective dose per dose length product- EDLP). Αρχικά η μελέτη επικεντρώθηκε σε νεογνά, ωστόσο εφόσον οι συντελεστές μετατροπής δόσης εξαρτώνται από τις διαστάσεις του ασθενούς, θεωρήθηκε σκόπιμο να επεκταθεί και σε εξετάσεις παιδιών, συμπεριλαμβανομένων των ηλικιών 1, 5 και 10 ετών.Η μέθοδος και τα αποτελέσματα αυτής της μελέτης μπορεί να παρέχουν ένα πρακτικό και αποτελεσματικό τρόπο στους ακτινολόγους και στους τεχνολόγους για την αποτελεσματική εκτίμηση της δόσης στην καθημερινή κλινική πρακτική. Αυτό μπορεί να συμβάλει στη βελτιστοποίηση των παραμέτρων έκθεσης και να βοηθήσει στον υπολογισμό της πιθανότητας εμφάνισης καρκίνου εξαιτίας της ακτινοβολίας για νεογνά και παιδιά διαφορετικών διαστάσεων, ενισχύοντας έτσι τα κριτήρια αιτιολόγησης της εξέτασης.

Conselho Nacional do Desenvolvimento Científico e Tecnológico (CNPq) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Financiadora de Estudos e Projetos (FINEP) / Fundação Araucária de Apoio ao Desenvolvimento Científico e Tecnológico do Paraná / Universidade Tecnológica Federal do Paraná (UTFPR) / Os sistemas de ultrassom para imagens médicas sempre estão em evolução. Na área de imagem Doppler, em que se pode observar o movimento do objeto, principalmente o fluxo sanguíneo, encontram-se algumas técnicas para obtenção de melhor qualidade de imagem. Os principais problemas da técnica Doppler são a complexidade e a quantidade de dados a serem processados para a formação das imagens. Este trabalho buscou avaliar dois métodos para processamento de imagens Doppler. Inicialmente, foram realizados estudos com o beamforming tradicional gerado pela plataforma de pesquisas ULTRA-ORS no laboratório de ultrassom. Após, foram realizados estudos com o equipamento Verasonics Vantage™ utilizando geração de onda de ultrassom planas, plane-wave, com um transdutor linear modelo L11-4v de 128 elementos. Os phantoms utilizados foram o CIRS Doppler String Phantom model 043 e o Doppler Flow Phantoms & Pumping Systems da ATS. Foram implementados algoritmos para reconstrução das imagens em modo B no Matlab® utilizando os métodos delay-and-sum e f-k migration, com a geração de imagens Doppler colorido e Power Doppler. Os dados para geração de imagem modo B com plane-wave foram adquiridos com variação de 1 a 75 ângulos, na faixa entre -8,88° e 8,88°, com passo de 0,24°. Os resultados obtidos com o f-k migration apresentaram maior resolução, com erros de 1,0 % e 0,8 % para as resoluções lateral e axial, respectivamente, enquanto o método DAS apresentou erros de 12,0 % para resolução lateral e 10,0 % para resolução axial. Para geração das imagens Doppler com plane-wave os dados foram adquiridos com variação de 1 a 21 ângulos, na faixa entre -15,0° e 15,0°, com passo de 1,5°. A estimação da velocidade através da técnica Doppler apresentou melhores resultados utilizando-se o método DAS, com erro de 8,3 %, enquanto o método f-k migration apresentou erro de 16,6 %. Analisando-se os resultados obtidos, foi possível verificar que a técnica plane-wave permite a geração de imagens com maior taxa de quadros por segundo do que os métodos tradicionais. Também pode se observar que o método f-k migration gera imagens de maior qualidade com menor número de ângulos de aquisição, cerca de 9 ângulos, porém possui pior desempenho para geração de imagens Doppler quando comparado ao DAS. / The medical ultrasound equipment is always evolving. In the field of Doppler imaging, which object movement and mainly blood flow of vessels can be measured, there are some techniques to improve image quality. The main problems of the Doppler technique are the complexity and the amount of data to be processed for the image reconstruction. The aim of this work was to evaluate two methods for Doppler images processing. Initially, studies were carried out with the traditional beamforming technique, generated by the research platform ULTRA-ORS in the ultrasound laboratory. Then, with the Verasonics Vantage™ equipment, it was generated ultrasound plane waves with a linear transducer L11-4v of 128 elements. Two Doppler phantoms were used, the CIRS Doppler String Phantom model 043 and the Doppler Flow Phantoms & Pumping Systems from ATS. Algorithms for B mode image reconstruction were developed in Matlab® using the methods Delay-and-Sum and f-k Migration to generate Color Doppler and Power Doppler images. The B mode images with plane-wave were generated from the data acquired with 1 to 75 angles, ranging from -8.88° to 8.88° and 0.24º step. The f-k migration’s results presented higher resolutions than DAS method, with errors of 1.0 % and 0.8 % for lateral and axial resolutions, respectively, while the DAS method presented errors of 12.0 % for lateral resolution and 10.0 % for axial resolution. The data for color Doppler images with plane-wave generation were acquired with 1 to 21 angles, ranging from -15.0° to 15.0°and 1.5° step. The Doppler velocity estimation using the DAS method showed better results (error of 8.3 %) than the f-k migration (error of 16.6 %). Analyzing the obtained results, it was possible to see that the plane wave imaging technique allows the improvement of the frame-rate, being faster than traditional methods. Additionally, it was verified that f-k migration method produces images with better quality using less steering angles, approximately 9 angles, but it shows worse performance when generating Doppler images.

A inspeção de defeitos de soldagem em imagens radiográficas de tubulações é bastante subjetiva e está sujeita a erros de interpretação por parte do inspetor laudista. Dentro desse contexto, nos últimos anos tem-se visto um grande esforço no desenvolvimento de métodos automáticos e semiautomáticos de detecção de defeitos em juntas soldadas. Este trabalho apresenta um método automatizado para detecção e classificação de defeitos em imagens radiográficas de juntas soldadas de tubulações obtidas pela técnica de exposição radiográfica parede dupla vista dupla (PDVD), obtidas em reais situações de campo e que, geralmente, têm uma qualidade mais baixa do que as imagens usadas em outros estudos. O método proposto identifica na imagem a região do cordão de solda, detecta as descontinuidades e classifica as mesmas em defeitos e não defeitos, destacando na imagem o resultado. São avaliados classificadores a partir de métodos de classificação por redes neurais Multilayer Perceptron (MLP), redes neurais Extreme Learning Machines (ELM) e classificador estatístico Support Vector Machines (SVM). O método proposto para identificação da região de interesse atingiu 100% de precisão na segmentação do cordão de solda. O classificador SVM apresentou um desempenho melhor que os classificadores MLP e ELM em todos os cenários testados. Com a utilização de ensembles de ELMs obteve-se um F-score de 85,7% para o banco de padrões de teste, resultados satisfatórios quando comprados com trabalhos semelhantes. O uso de ensembles de ELMs representa um ganho de apenas 0,5% no F-score em comparação com o melhor resultado de rede treinada individualmente, entretanto, com o uso de faixas de limiares de decisão do ensemble, o uso do método permite mostrar as descontinuidades sobre as quais o ensemble não tem certeza, destacando na imagem estas descontinuidades. A imagem resultate da aplicação do método serve como auxílio ao especialista na elaboração de laudos. / The inspection of radiographic images of welded joints is very subjective and is subject to errors of interpretation by the inspector. In this context, a great effort has been made in the last years to develop automatic and semiautomatic methods for detecting defects in welded joints. This research work presents an automated method for the detection and classification of defects in radiographic images of welded joints of pipes obtained by the double wall double image (DWDI) exposure technique obtained in real field situations and which generally have a lower quality than the images used in other studies. The proposed methos identifies the region of the weld bead, detects the discontinuities and classifies them as defects and non-defects, highlighting in the image the result. Classifiers are evalueted using methods of classification by multilayer perceptron (MLP) neural networks, extreme learning machines (ELM) neural networks, and Support Vector Machines (SVM). The proposed method for identifying the region of interest reached 100% precision in the segmentation od the weld bead. The SVM classifier performed better than the MLP and ELM classifiers in all scenarios tested. Using ELM ensembles, an F_score of 85,7% was obtained for a test patterns database, satisfactoryresults when compared to similar works. The use of ensembles of ELMs represents a gain of only 0,5% in the F-score compared to the best result of the individually trained network, however, with the use of ensemble decision threshold ranges, the presented method allows to show the discontinuities about which the ensemble is not sure, highlighting in the image these discontinuities as a region of uncertainty, leaving to the specialist the final evaluation of these discontinuities. The image resulting from the application of the method serves as an aid to the expert in the elaboration of reports.

A avaliação de qualidade de vídeo possui um papel fundamental no processamento de vídeo e em aplicações de comunicação. Uma métrica de qualidade de vídeo ideal deve garantir a alta correlação entre a predição da distorção do vídeo e a percepção de qualidade do Sistema Visual Humano. Este trabalho propõe o uso de modelos de atenção visual com abordagem bottom up baseados em saliências para avaliação de qualidade de vídeo. Três métricas objetivas de avaliação são propostas. O primeiro método é uma métrica com referência completa baseada na estrutura de similaridade. O segundo modelo é uma métrica sem referência baseada em uma modelagem sigmoidal com solução de mínimos quadrados que usa o algoritmo de Levenberg-Marquardt e extração de características espaço-temporais. E, a terceira métrica é análoga à segunda, porém usa a característica Blockiness na detecção de distorções de blocagem no vídeo. A abordagem bottom-up é utilizada para obter os mapas de saliências que são extraídos através de um modelo multiescala de background baseado na detecção de movimentos. Os resultados experimentais apresentam um aumento da eficiência de predição de qualidade de vídeo nas métricas que utilizam o modelo de saliência em comparação com as respectivas métricas que não usam este modelo, com destaque para as métricas sem referência propostas que apresentaram resultados melhores do que métricas com referência para algumas categorias de vídeos. / Video quality assessment plays a key role in the video processing and communications applications. An ideal video quality metric shall ensure high correlation between the video distortion prediction and the perception of the Human Visual System. This work proposes the use of visual attention models with bottom-up approach based on saliencies for video qualitty assessment. Three objective metrics are proposed. The first method is a full reference metric based on the structural similarity. The second is a no reference metric based on a sigmoidal model with least squares solution using the Levenberg-Marquardt algorithm and extraction of spatial and temporal features. And, the third is analagous to the last one, but uses the characteristic Blockiness for detecting blocking distortions in the video. The bottom-up approach is used to obtain the salient maps, which are extracted using a multiscale background model based on motion detection. The experimental results show an increase of efficiency in the quality prediction of the proposed metrics using salient model in comparission to the same metrics not using these model, highlighting the no reference proposed metrics that had better results than metrics with reference to some categories of videos.

No Brasil, o câncer de mama é o mais frequente e a principal causa de morte entre as mulheres, para o ano de 2016 foram estimados 57.960 mil novos casos. A Tomossíntese Digital Mamária (TDM) reduz os efeitos da sobreposição do tecido da mama, oferecendo uma melhor caracterização dos achados mamográficos e diminuindo a necessidade de incidências adicionais. A recente introdução da TDM na prática clínica requer manutenção dos requisitos de dose e qualidade da imagem, além de rotinas de Controle de Qualidade (CQ) e monitoração da Dose Glandular Média (DGM), e também uma equipe técnica qualificada para executar o procedimento. A prática de monitoração da qualidade da imagem e da dose de radiação é uma solução para assegurar continuamente a alta qualidade de exames de radiodiagnóstico. São poucos os estudos relacionados ao CQ em TDM no Brasil, e ainda não foram estabelecidos protocolos nacionais e internacionais específicos; sendo assim, os testes de CQ utilizados são baseados naqueles utilizados pela National Health Service Breast Screening Programme (NHSBSP), European Reference Organisation for Quality Assured Breast Screening and Diagnostic Service (EUREF), American Association of Physicists in Medicine (AAPM), Internacional Atomic Energy Agency (IAEA). Este estudo propõe e apresenta os resultados dos testes de CQ e DGM para mamografia e TDM, e um Protocolo de Rotina para exames de TDM. Os testes de CQ para TDM foram realizados conforme as recomendações do fabricante, da Portaria 2898/13 e de protocolos internacionais. O Protocolo de Rotina foi elaborado visando sanar as dúvidas mais frequentes da equipe técnica, assim como otimizar a qualidade do exame em sala. Para atestar a qualidade da imagem diagnóstica para todas as espessuras, foi realizado o teste de Qualidade de Imagem do simulador padrão do American College of Radiology (ACR). Na avaliação da DGM todos os valores obtidos para mamografia e TDM se mantiveram em conformidade com os limites de referência utilizados, exceto para a espessura de 20 mm em TDM. Para todos os outros testes de CQ, os resultados se mantiveram em conformidade com os limites de referência. Os resultados adquiridos nas avaliações reforçam a importância do controle de qualidade aplicado de forma adequada no processo de obtenção das imagens, assegurando baixa dose e boa qualidade da imagem, cumprindo o princípio ALARA. / In Brazil, breast cancer is the most common and the leading cause of death among women; in 2015, 57,120,000 new cases. Digital Breast Tomosynthesis (DBT) reduces the effects of overlapping breast tissue, offering a better characterization of mammographic findings and decreasing the need of additional views. The recent introduction of DBT in clinical practice requires, in order to maintain the dose requirements and image quality, aside from routine Quality Control (QC) procedures and monitoring the Mean Glandular Dose (MGD) also a qualified technical team to perform the procedure. The practice of monitoring the image quality and the radiation dose is a solution to continuously ensure high quality diagnostic radiology exams. Few studies related to CQ TDM in Brazil, and national and international specific protocols have not been established, Thus, QC tests used are based on those used by the National Health Service Breast Screening Programme (NHSBSP), European Reference Organization for Quality Assured Breast Screening and Diagnostic Service (EUREF), American Association of Physicists in Medicine (AAPM), International Atomic Energy Agency (IAEA). This study proposes and presents the results of the QC tests and DGM for mammography and TDM, and routine protocol for TDM tests. The QC tests for TDM were performed according to the manufacturer's recommendations, of Portaria 2898/13 and international protocols. The Routine Protocol was elaborated aiming to resolve the most frequent doubts of the technical team, as well as optimize the room examination. In order to certify the quality of the diagnostic image for all thicknesses, a imagine quality test from the standard simulator of American College of Radiology (ACR) was performed. In the DGM evaluation, all values mammography and TDM remained in accordance with the reference range used except for 20 mm TDM. For all other QC tests, the results remained in accordance with the reference limits. The results obtained in the assessments emphasize the importance of quality control applied properly in obtaining the imaging process, ensuring low dose and good image quality, fulfilling the ALARA principle.

A avaliação de qualidade de vídeo possui um papel fundamental no processamento de vídeo e em aplicações de comunicação. Uma métrica de qualidade de vídeo ideal deve garantir a alta correlação entre a predição da distorção do vídeo e a percepção de qualidade do Sistema Visual Humano. Este trabalho propõe o uso de modelos de atenção visual com abordagem bottom up baseados em saliências para avaliação de qualidade de vídeo. Três métricas objetivas de avaliação são propostas. O primeiro método é uma métrica com referência completa baseada na estrutura de similaridade. O segundo modelo é uma métrica sem referência baseada em uma modelagem sigmoidal com solução de mínimos quadrados que usa o algoritmo de Levenberg-Marquardt e extração de características espaço-temporais. E, a terceira métrica é análoga à segunda, porém usa a característica Blockiness na detecção de distorções de blocagem no vídeo. A abordagem bottom-up é utilizada para obter os mapas de saliências que são extraídos através de um modelo multiescala de background baseado na detecção de movimentos. Os resultados experimentais apresentam um aumento da eficiência de predição de qualidade de vídeo nas métricas que utilizam o modelo de saliência em comparação com as respectivas métricas que não usam este modelo, com destaque para as métricas sem referência propostas que apresentaram resultados melhores do que métricas com referência para algumas categorias de vídeos. / Video quality assessment plays a key role in the video processing and communications applications. An ideal video quality metric shall ensure high correlation between the video distortion prediction and the perception of the Human Visual System. This work proposes the use of visual attention models with bottom-up approach based on saliencies for video qualitty assessment. Three objective metrics are proposed. The first method is a full reference metric based on the structural similarity. The second is a no reference metric based on a sigmoidal model with least squares solution using the Levenberg-Marquardt algorithm and extraction of spatial and temporal features. And, the third is analagous to the last one, but uses the characteristic Blockiness for detecting blocking distortions in the video. The bottom-up approach is used to obtain the salient maps, which are extracted using a multiscale background model based on motion detection. The experimental results show an increase of efficiency in the quality prediction of the proposed metrics using salient model in comparission to the same metrics not using these model, highlighting the no reference proposed metrics that had better results than metrics with reference to some categories of videos.

No Brasil, o câncer de mama é o mais frequente e a principal causa de morte entre as mulheres, para o ano de 2016 foram estimados 57.960 mil novos casos. A Tomossíntese Digital Mamária (TDM) reduz os efeitos da sobreposição do tecido da mama, oferecendo uma melhor caracterização dos achados mamográficos e diminuindo a necessidade de incidências adicionais. A recente introdução da TDM na prática clínica requer manutenção dos requisitos de dose e qualidade da imagem, além de rotinas de Controle de Qualidade (CQ) e monitoração da Dose Glandular Média (DGM), e também uma equipe técnica qualificada para executar o procedimento. A prática de monitoração da qualidade da imagem e da dose de radiação é uma solução para assegurar continuamente a alta qualidade de exames de radiodiagnóstico. São poucos os estudos relacionados ao CQ em TDM no Brasil, e ainda não foram estabelecidos protocolos nacionais e internacionais específicos; sendo assim, os testes de CQ utilizados são baseados naqueles utilizados pela National Health Service Breast Screening Programme (NHSBSP), European Reference Organisation for Quality Assured Breast Screening and Diagnostic Service (EUREF), American Association of Physicists in Medicine (AAPM), Internacional Atomic Energy Agency (IAEA). Este estudo propõe e apresenta os resultados dos testes de CQ e DGM para mamografia e TDM, e um Protocolo de Rotina para exames de TDM. Os testes de CQ para TDM foram realizados conforme as recomendações do fabricante, da Portaria 2898/13 e de protocolos internacionais. O Protocolo de Rotina foi elaborado visando sanar as dúvidas mais frequentes da equipe técnica, assim como otimizar a qualidade do exame em sala. Para atestar a qualidade da imagem diagnóstica para todas as espessuras, foi realizado o teste de Qualidade de Imagem do simulador padrão do American College of Radiology (ACR). Na avaliação da DGM todos os valores obtidos para mamografia e TDM se mantiveram em conformidade com os limites de referência utilizados, exceto para a espessura de 20 mm em TDM. Para todos os outros testes de CQ, os resultados se mantiveram em conformidade com os limites de referência. Os resultados adquiridos nas avaliações reforçam a importância do controle de qualidade aplicado de forma adequada no processo de obtenção das imagens, assegurando baixa dose e boa qualidade da imagem, cumprindo o princípio ALARA. / In Brazil, breast cancer is the most common and the leading cause of death among women; in 2015, 57,120,000 new cases. Digital Breast Tomosynthesis (DBT) reduces the effects of overlapping breast tissue, offering a better characterization of mammographic findings and decreasing the need of additional views. The recent introduction of DBT in clinical practice requires, in order to maintain the dose requirements and image quality, aside from routine Quality Control (QC) procedures and monitoring the Mean Glandular Dose (MGD) also a qualified technical team to perform the procedure. The practice of monitoring the image quality and the radiation dose is a solution to continuously ensure high quality diagnostic radiology exams. Few studies related to CQ TDM in Brazil, and national and international specific protocols have not been established, Thus, QC tests used are based on those used by the National Health Service Breast Screening Programme (NHSBSP), European Reference Organization for Quality Assured Breast Screening and Diagnostic Service (EUREF), American Association of Physicists in Medicine (AAPM), International Atomic Energy Agency (IAEA). This study proposes and presents the results of the QC tests and DGM for mammography and TDM, and routine protocol for TDM tests. The QC tests for TDM were performed according to the manufacturer's recommendations, of Portaria 2898/13 and international protocols. The Routine Protocol was elaborated aiming to resolve the most frequent doubts of the technical team, as well as optimize the room examination. In order to certify the quality of the diagnostic image for all thicknesses, a imagine quality test from the standard simulator of American College of Radiology (ACR) was performed. In the DGM evaluation, all values mammography and TDM remained in accordance with the reference range used except for 20 mm TDM. For all other QC tests, the results remained in accordance with the reference limits. The results obtained in the assessments emphasize the importance of quality control applied properly in obtaining the imaging process, ensuring low dose and good image quality, fulfilling the ALARA principle.