Improving the knowledge of the spectral and temporal properties of plasma-based XUV lasers is an important issue for the ongoing development of these sources towards significantly higher peak power. The spectral properties of the XUV laser line actually control several physical quantities that are important for applications, such as the minimum duration that can be achieved (Fourier-transform limit). The shortest duration experimentally achieved to-date is ~1 picosecond. The demonstrated technique of seeding XUV laser plasmas with a coherent femtosecond pulse of high-order harmonic radiation opens new and promising prospects to reduce the duration to a few 100 fs, provided that the gain bandwidth can be kept large enough.XUV lasers pumped by collisional excitation of Ni-like and Ne-like ions have been developed worldwide in hot plasmas created either by fast electrical discharge, or by various types of high-power lasers. This leads to a variety of XUV laser sources with distinct output properties, but also markedly different plasma parameters (density, temperature) in the amplification zone. Hence different spectral properties are expected. The purpose of our work was then to investigate the spectral behaviour of the different types of existing collisional excitation XUV lasers, and to evaluate their potential to support amplification of pulses with duration below 1 ps in a seeded mode.The spectral characterization of plasma-based XUV lasers is challenging because the extremely narrow bandwidth (typically ∆λ/λ ~10-5) lies beyond the resolution limit of existing spectrometers in this spectral range. In our work the narrow linewidth was resolved using a wavefront-division interferometer specifically designed to measure temporal coherence, from which the spectral linewidth is inferred. We have characterized three types of collisional XUV lasers, developed in three different laboratories: transient pumping in Ni-like Mo, capillary discharge pumping in Ne-like Ar and quasi-steady state pumping in Ne-like Zn. Besides the accurate measurement of the temporal coherence of the laser in each case, we have studied the spectral behaviour when the laser is operated in the saturation regime and (in Ni-like Mo) when it is seeded with high-order harmonic radiation. We have also investigated the temporal behaviour of the Ni-like Mo transient XUV laser, using an ultrafast X-ray streak camera. Our linewidth measurements are compared with detailed numerical calculations including relevant broadening mechanisms as well as radiative transfer effects. The evolution of the spectral profile with amplification and saturation was studied for different plasma parameters, and corresponding Fourier-transform limit duration were evaluated.The shortest temporal coherence (ie the largest bandwidth) is measured for the quasi-steady state pumping XUV laser, which operates at the highest density and ionic temperature.

<p>Temporal processing and temporal coherence processing are fundamental components of auditory processing and the focus of this thesis. Cortical temporal processing in particular is understudied in humans. This dissertation makes three contributions that help characterize auditory temporal processing, and its relationship to auditory perception in humans. Experiment 1 develops a novel systems identification approach utilizing modified maximum length sequences (m-seq) to robustly measure cortical temporal processing noninvasively. Using this technique, it is found that cortex's ability to track dynamic spatial auditory cues can explain the ability to utilize dynamic binaural information to do a spatial unmasking task. This result combined with behavioral data that shows FM tracking extends out to similar rates as spatial tracking demonstrates how auditory tracking broadly is constrained by the temporal properties of cortex. Experiment 2 develops a novel measure of temporal processing, called the modulation temporal response function (mod-TRF), which can separate the underlying sources contributing to modulation processing along the auditory system from short, middle, and late latency regions. The mod-TRF has a robust SNR at the individual level giving it the potential to become an ubiquitous tool to assess temporal processing and auditory activity generally across individuals. The utility of the mod-TRF is demonstrated by evaluating how attention affects different sources along the auditory pathway. Other studies utilizing the mod-TRF could explore how temporal processing in early and late areas of the auditory system changes with aging, hearing loss, musicianship, and neuropsychiatric disorders. Experiment 3 establishes a novel physiological measure of across channel temporal coherence processing to measure auditory binding ability across individuals. It  is demonstrated how this novel temporal coherence measure can explain performance on a behavioral temporal coherence detection task, speech-in-noise task, and comodulation masking release. Lastly, it is explored how social and communicative features in individuals, measured via the Autism Quotient, align with differences in auditory ability at various tasks. </p>

No description available.

La caractérisation spectrale détaillée des lasers XUV générés dans des plasmas est un enjeu important des projets actuels de développement qui visent à augmenter la puissance crête de ces sources. En effet les propriétés spectrales de ces lasers conditionnent d'autres propriétés importantes, telle que la durée minimum accessible (limite de Fourier). La plus courte durée actuellement démontrée expérimentalement est de 1 picoseconde. La technique d'injection d'un plasma de laser XUV avec une impulsion femtoseconde de rayonnement harmonique d'ordre élevé offre des perspectives très prometteuses pour réduire la durée d'impulsion jusqu'à quelques 100 femtosecondes, pourvu que l'on sache maintenir une bande spectrale de gain suffisamment large.Les lasers XUV pompés par excitation collisionnelle dans des ions néonoïdes et nickeloïdes ont été développés dans des plasmas chauds créés aussi bien par décharge électrique rapide que par différents types de lasers de puissance. On a ainsi accès à une large variété de sources lasers XUV, qui diffèrent par les caractéristiques du faisceau émis, mais aussi par les paramètres du plasma (densité, température) dans la zone de gain. On peut donc s'attendre à des propriétés spectrales différentes. Le but du travail que nous présentons est d'étudier les propriétés spectrales des différents types de lasers XUV collisionnels existants, et d'évaluer leur capacité à amplifier des impulsions de durée inférieure à 1 picoseconde, dans un mode injecté.La caractérisation spectrale des lasers XUV est expérimentalement difficile parce que la résolution spectrale nécessaire (∆λ/λ ~10-5) n'est pas accessible avec les meilleurs spectromètres actuels. Dans notre étude, nous avons atteint cette résolution en mesurant la cohérence temporelle de la source à l'aide d'un interféromètre à division de front d'onde, spécifiquement conçu pour ces mesures, à partir desquelles largeur spectrale peut être déduite.Nous avons caractérisé trois types de lasers XUV collisionnels, développés dans trois laboratoires différents: pompage transitoire dans le molybdène nickeloïde, pompage par décharge électrique dans l'argon néonoïde et pompage quasi-stationnaire dans le zinc néonoïde. Dans chaque cas la cohérence temporelle a été mesurée précisément. De plus nous avons étudié l'effet de la saturation de l'amplification et (pour le Ni-like Mo) l'influence du mode injecté. Nous avons également étudié le comportement temporel du laser transitoire Ni-like Mo à l'aide d'une caméra streak X ultra-rapide. Nos mesures spectrales sont comparées à des résultats de simulations numériques prenant en compte les différents mécanismes d'élargissement ainsi que les effets de transfert radiatif. Nous avons étudié l'évolution du profil spectral avec l'amplification et la saturation, et nous avons évalué les limites de Fourier correspondantes.Le temps de cohérence le plus court (ie la largeur spectrale la plus grande) est mesuré pour le laser XUV quasi-stationnaire, qui correspond au plasma qui a la plus forte densité et la plus forte température ionique. / Improving the knowledge of the spectral and temporal properties of plasma-based XUV lasers is an important issue for the ongoing development of these sources towards significantly higher peak power. The spectral properties of the XUV laser line actually control several physical quantities that are important for applications, such as the minimum duration that can be achieved (Fourier-transform limit). The shortest duration experimentally achieved to-date is ~1 picosecond. The demonstrated technique of seeding XUV laser plasmas with a coherent femtosecond pulse of high-order harmonic radiation opens new and promising prospects to reduce the duration to a few 100 fs, provided that the gain bandwidth can be kept large enough.XUV lasers pumped by collisional excitation of Ni-like and Ne-like ions have been developed worldwide in hot plasmas created either by fast electrical discharge, or by various types of high-power lasers. This leads to a variety of XUV laser sources with distinct output properties, but also markedly different plasma parameters (density, temperature) in the amplification zone. Hence different spectral properties are expected. The purpose of our work was then to investigate the spectral behaviour of the different types of existing collisional excitation XUV lasers, and to evaluate their potential to support amplification of pulses with duration below 1 ps in a seeded mode.The spectral characterization of plasma-based XUV lasers is challenging because the extremely narrow bandwidth (typically ∆λ/λ ~10-5) lies beyond the resolution limit of existing spectrometers in this spectral range. In our work the narrow linewidth was resolved using a wavefront-division interferometer specifically designed to measure temporal coherence, from which the spectral linewidth is inferred. We have characterized three types of collisional XUV lasers, developed in three different laboratories: transient pumping in Ni-like Mo, capillary discharge pumping in Ne-like Ar and quasi-steady state pumping in Ne-like Zn. Besides the accurate measurement of the temporal coherence of the laser in each case, we have studied the spectral behaviour when the laser is operated in the saturation regime and (in Ni-like Mo) when it is seeded with high-order harmonic radiation. We have also investigated the temporal behaviour of the Ni-like Mo transient XUV laser, using an ultrafast X-ray streak camera. Our linewidth measurements are compared with detailed numerical calculations including relevant broadening mechanisms as well as radiative transfer effects. The evolution of the spectral profile with amplification and saturation was studied for different plasma parameters, and corresponding Fourier-transform limit duration were evaluated.The shortest temporal coherence (ie the largest bandwidth) is measured for the quasi-steady state pumping XUV laser, which operates at the highest density and ionic temperature.

In recent times there has been an explosion of online user-generated video content. This has generated significant research interest in video analytics. Human users understand videos based on high-level semantic concepts. However, most of the current research in video analytics are driven by low-level features and descriptors, which often lack semantic interpretation. Existing attempts in semantic video analytics are specialized and require additional resources like movie scripts, which are not available for most user-generated videos. There are no general purpose approaches to understanding videos through semantic concepts. In this thesis we attempt to bridge this gap. We view videos as collections of entities which are semantic visual concepts like the persons in a movie, or cars in a F1 race video. We focus on two fundamental tasks in Video Understanding, namely summarization and scene- discovery. Entity-driven Video Summarization and Entity-driven Scene discovery are important open problems. They are challenging due to the spatio-temporal nature of videos, and also due to lack of apriori information about entities. We use Bayesian nonparametric methods to solve these problems. In the absence of external resources like scripts we utilize fundamental structural properties like temporal coherence in videos- which means that adjacent frames should contain the same set of entities and have similar visual features. There have been no focussed attempts to model this important property. This thesis makes several contributions in Computer Vision and Bayesian nonparametrics by addressing Entity-driven Video Understanding through temporal coherence modeling. Temporal Coherence in videos is observed across its frames at the level of features/descriptors, as also at semantic level. We start with an attempt to model TC at the level of features/descriptors. A tracklet is a spatio-temporal fragment of a video- a set of spatial regions in a short sequence (5-20) of consecutive frames, each of which enclose a particular entity. We attempt to find a representation of tracklets to aid tracking of entities. We explore region descriptors like Covari- ance Matrices of spatial features in individual frames. Due to temporal coherence, such matrices from corresponding spatial regions in successive frames have nearly identical eigenvectors. We utilize this property to model a tracklet using a covariance matrix, and use it for region-based entity tracking. We propose a new method to estimate such a matrix. Our method is found to be much more efficient and effective than alternative covariance-based methods for entity tracking. Next, we move to modeling temporal coherence at a semantic level, with special emphasis on videos of movies and TV-series episodes. Each tracklet is associated with an entity (say a particular person). Spatio-temporally close but non-overlapping tracklets are likely to belong to the same entity, while tracklets that overlap in time can never belong to the same entity. Our aim is to cluster the tracklets based on the entities associated with them, with the goal of discovering the entities in a video along with all their occurrences. We argue that Bayesian Nonparametrics is the most convenient way for this task. We propose a temporally coherent version of Chinese Restaurant Process (TC-CRP) that can encode such constraints easily, and results in discovery of pure clusters of tracklets, and also filter out tracklets resulting from false detections. TC-CRP shows excellent performance on person discovery from TV-series videos. We also discuss semantic video summarization, based on entity discovery. Next, we consider entity-driven temporal segmentation of a video into scenes, where each scene is characterized by the entities present in it. This is a novel application, as existing work on temporal segmentation have focussed on low-level features of frames, rather than entities. We propose EntScene: a generative model for videos based on entities and scenes, and propose an inference algorithm based on Blocked Gibbs Sampling, for simultaneous entity discovery and scene discovery. We compare it to alternative inference algorithms, and show significant improvements in terms of segmentatio and scene discovery. Video representation by low-rank matrix has gained popularity recently, and has been used for various tasks in Computer Vision. In such a representation, each column corresponds to a frame or a single detection. Such matrices are likely to have contiguous sets of identical columns due to temporal coherence, and hence they should be low-rank. However, we discover that none of the existing low-rank matrix recovery algorithms are able to preserve such structures. We study regularizers to encourage these structures for low-rank matrix recovery through convex optimization, but note that TC-CRP-like Bayesian modeling is better for enforcing them. We then focus our attention on modeling temporal coherence in hierarchically grouped sequential data, such as word-tokens grouped into sentences, paragraphs, documents etc in a text corpus. We attempt Bayesian modeling for such data, with application to multi-layer segmentation. We first make a detailed study of existing models for such data. We present a taxonomy for such models called Degree-of-Sharing (DoS), based on how various mixture components are shared by the groups of data in these models. We come up with Layered Dirichlet Process which generalizes Hierarchical Dirichlet Process to multiple layers, and can also handle sequential information easily through Markovian approach. This is applied to hierarchical co-segmentation of a set of news transcripts- into broad categories (like politics, sports etc) and individual stories. We also propose a explicit-duration (semi-Markov) approach for this purpose, and provide an efficient inference algorithm for this. We also discuss generative processes for distribution matrices, where each column is a probability distribution. For this we discuss an application: to infer the correct answers to questions on online answering forums from opinions provided by different users.

La technique de l'interférométrie en astronomie permet d'observer des objets avec une haute résolution angulaire comparativement à l'utilisation d'un unique télescope. L'observation dans l'infrarouge moyen représente aujourd'hui un enjeu en interférométrie notamment pour l'étude des noyaux actifs de galaxie ou de la formation des planètes. Cependant ce domaine spectral est particulièrement contraignant puisqu'il est soumis à l'émission propre des éléments optiques de l'instrument mais également de l'atmosphère. Ce manuscrit développe les travaux effectués sur un nouvel instrument utilisant un processus de conversion de fréquence pour transposer le rayonnement infrarouge vers un domaine spectral permettant de s'affranchir de l'émission propres des optiques. Un prototype fonctionnant dans le proche infrarouge à 1,55 µm et convertissant, via une processus de somme de fréquences, le rayonnement dans le domaine visible autour de 630 nm grâce à une pompe intense à 1064 nm, a été mis en place pour démontrer, en laboratoire, le principe de cette solution innovante notamment dans le cadre de l'analyse de la cohérence spatiale d'un corps noir. L'objectif est maintenant de démontrer la capacité de l'instrument à détecter un objet réel. J'introduis dans cette thèse les notions théoriques essentielles à la compréhension des travaux présentés pour ensuite détailler le fonctionnement de l'instrument et les éléments d'amélioration apportés, notamment en terme de transmission, au cours de ma thèse. Les études préliminaires en laboratoire du comportement de l'instrument ont permis d'aboutir aux premières franges sur le ciel en utilisant la plus petite base (34 m) du réseau interférométrique CHARA et de rechercher la magnitude limite de l'instrument. L'utilisation du processus de conversion de fréquence a pour conséquence de filtrer le spectre converti. Ainsi dans la configuration actuelle de l'interféromètre, seul 0,6 nm du spectre infrarouge en entrée du cristal est converti à travers le processus de SFG. Afin d'augmenter la sensibilité, une solution est de créer plusieurs processus de SFG simultanément dans chaque étage de conversion afin d'échantillonner le spectre infrarouge converti. Cette solution requiert d'utiliser plusieurs sources de pompe indépendantes qui vont créer des systèmes de franges incohérents. Je présente l'analyse de la cohérence temporelle d'une source infrarouge large bande convertie via l'utilisation de deux sources de pompe et un moyen de synchroniser les différents systèmes de franges afin de maximiser le contraste. / Interferometry is an instrumental technique suitable to perform astronomical observations at high angular resolution. Currently, the mid-infrared spectral domain is a real issue for the astronomical interferometry to characterize astronomical objects such as proto-planetary discs or active galactic nuclei. However, this spectral domain is subject to a large thermal background emission from the instrument and from the sky. This manuscript describes an innovative instrument using a nonlinear process of sum frequency generation to convert the mid-infrared radiation to a shorter wavelength domain where the thermal emission from the instrument is negligible. A prototype operating in the near-infrared at 1.55 µm and converting the radiation in the visible domain at 630 nm thanks to a strong pump at 1064 nm has already demonstrated its ability to analyse spatial coherence of a blackbody source. The present goal is to demonstrate its ability to detect an object on the sky. In this manuscript I introduce theoritical concepts necessary for an understanding of the overall operation of the instrument. Then, I describe the main improvements provided in this thesis, in particular concerning the instrumental transmission. The preliminary studies of the instrument operation resulted in the first on-sky fringes on the CHARA array and enabled to determine its limiting magnitude. The upconversion process acts as a filter on the converted spectrum. In the current instrumental configuration, only 0.6 nm of the input infrared spectrum is converted through the SFG process. For the purpose of increasing the instrumental sensitivity, we propose to sample the infrared spectrum by using several independent pump laser lines thus creating different incoherent fringe patterns. I present the temporal coherence analysis of a broadband infrared source converted by a dual-line pump laser and a method to synchronize the different fringe patterns to insure a maximum value of the contrast.

Super-resolution is a subfield of computer vision centered around upsampling low-resolution images to a corresponding high-resolution counterpart. This degree project investigates the suitability of a deep learning method for real-time video super-resolution. Following earlier work in the field, we use bicubic interpolation as a baseline for comparison. The deep learning method selected is specifically suited towards real-time super-resolution and consists of a motion compensation network and an upsampling network. The deep learning method and bicubic interpolation are compared by quantitatively evaluating the methods against each other in quality metrics and performance metrics. Suitable quality metrics are selected from earlier works to provide increased comparability of results, namely peak signal-to-noise ratio and structure similarity index. The performance metrics are: number of operations for a single upsampled frame, latency, throughput, and memory requirements. We apply the methods to a highly challenging publicly available dataset specifically engineered towards video super-resolution research. To further investigate the deep learning method, we propose a few modifications and study the effect on the metrics. Our findings show that the deep learning models outperform bicubic interpolation in the quality metrics, while bicubic interpolation outperformed the deep learning models in the performance metrics. We also find no significant quality metric improvement associated with having a motion compensation network for this dataset, suggesting that the dataset might be too complex for the motion compensation network. We conclude that the deep learning method exhibits real-time capabilities as the method has a throughput of around 500 frames per second for full HD super-resolution. Additionally, we show that by modifying the deep learning method, we achieve similar latency as bicubic interpolation without sacrificing throughput or quality. / Superupplösning är ett område inom datorseende centrerat kring att uppsampla lågupplösta bilder till högupplösta motsvarigheter. Detta examensarbete undersöker hur lämplig en specifik djupinlärningsmetod är för superupplösning i realtid. Enligt tidigare forskning använder vi oss av bikubisk interpolering som grund för jämförelse. Den valda djupinlärningsmetoden är speciellt anpassad till superupplösning i realtid och består av ett rörelsekompensationsnätverk och ett uppsamplingsnätverk. Djupainlärningsmetoden och interpoleringsmetoden jämförs genom att kvantitativt utvärdera metoderna mot varandra i kvalitetsmått och prestandamått. Lämpliga kvalitetsmått väljs från tidigare forskning för att ge ökad jämförbarhet mellan resultaten, nämligen maximalt signaltill- brusförhållande och strukturlikhetsindex. Prestandamätvärdena är: antal operationer för en uppsamplad bild, latens, genomströmning och minnesbehov. Vi utvärderar metoderna på ett utmanande allmänt tillgängligt dataset speciellt konstruerat för superupplösningsforskning inom video. För att ytterligare undersöka den djupa inlärningsmetoden föreslår vi några modifieringar och studerar effekten på mätvärdena. Våra resultat visar att djupinlärningsmodellerna överträffar bikubisk interpolering i kvalitetsmåtten, medan bikubisk interpolering överträffar djupinlärningsmodellerna i prestandamåtten. Vi finner inte heller någon signifikant kvalitetsmässig förbättring förknippad med att ha ett rörelsekompensationsnätverk för detta dataset, vilket kan betyda att datasetet är för komplext för rörelsekompensationnätverket. Vi drar slutsatsen att djupainlärningsmetoden uppvisar realtidsfunktioner eftersom metoden har en genomströmning på cirka 500 bilder per sekund för full HD superupplösning. Dessutom visar vi att genom att modifiera djupainlärningsmetoden uppnår vi liknande latens som bikubisk interpolering utan att offra genomströmning eller kvalitet.

Le travail présenté dans cette thèse repose sur l'expertise du groupe VeCSEL du l'Institut d'Électronique de Sud (IES) dans la conception, la réalisation et l'étude des VeCSEL dans le proche et le moyen infrarouge. Avant de s'engager dans le travail présenté ici, les VeCSELs développés au sein de notre laboratoire concernaient l'habituel fonctionnement monofréquence avec un faisceau TEM00, mais repoussé à forte puissance et à haute cohérence. L'objectif de cette thèse est d'aller au delà de cet état conventionnel, et explore d'autres états cohérents du photon, inhabituels dans le domaine de l'émission laser. Ces nouveaux états concernent aussi bien l'aspect transverse (spatial) que longitudinal (temporel) de l'onde. Plus particulièrement, l'émission laser de haute cohérence et de forte puissance sur les modes d'ordre supérieur Laguerre-Gauss dégénérés et non-dégénérés (vortex optiques), et les lasers large bande cohérents appelé aussi «~laser sans mode~». / The work presented in this thesis is based on the expertise of the VECSEL group of the Institut d'Électronique de Sud (IES) in the design, implementation and study of VECSEL in the near and mid infrared. Before engaging in the work presented here, the VeCSELs developed in our laboratory involved the usual single-frequency operation with a TEM 00 beam, but pushed to high power and high coherence. The objective of this thesis is to go beyond this conventional state and to explore other coherent states of the photon, unusual in the field of laser emission. These new states apply to both the transverse appearance (spatial) and longitudinal (time) of the wave. More specifically, the laser emission of high power and high coherence on the Laguerre-Gauss higher order modes degenerate and non-degenerate (optical vortex), and the wide band coherent lasers also called "modeless lasers".

<pre><p>Difficulty understanding speech in background noise is the most common hearing complaint. Elucidating the neurophysiological mechanisms underlying speech intelligibility in everyday environments with multiple sound sources and distortions is hence important for any technology that aims to improve real-world listening. Using a combination of behavioral, electroencephalography (EEG), and computational modeling experiments, this dissertation provides insight into how the brain analyzes such complex scenes, and what roles different acoustic cues play in facilitating this process and in conveying phonetic content. Experiment #1 showed that brain oscillations selectively track the temporal envelopes (i.e., modulations) of attended speech in a mixture of competing talkers, and that the strength and pattern of this attention effect differs between individuals. Experiment #2 showed that the fidelity of neural tracking of attended-speech envelopes is strongly shaped by the modulations in interfering sounds as well as the temporal fine structure (TFS) conveyed by the cochlea, and predicts speech intelligibility in diverse listening environments. Results from Experiments #1 and #2 support the theory that temporal coherence of sound elements across envelopes and/or TFS shapes scene analysis and speech intelligibility. Experiment #3 tested this theory further by measuring and computationally modeling consonant categorization behavior in a range of background noises and distortions. We found that a physiologically plausible model that incorporated temporal-coherence effects predicted consonant confusions better than conventional speech-intelligibility models, providing independent evidence that temporal coherence influences scene analysis. Finally, results from Experiment #3 also showed that TFS is used to extract speech content (voicing) for consonant categorization even when intact envelope cues are available. Together, the novel insights provided by our results can guide future models of speech intelligibility and scene analysis, clinical diagnostics, improved assistive listening devices, and other audio technologies.</p></pre>