No description available.

No description available.

Accurate models of visual system are key for understanding how our brains process visual information. In recent years, deep neural networks (DNN) have been rapidly gaining traction in this domain. However, only few studies attempted to incorporate known anatomical properties of visual system into standard DNN architectures adapted from the general machine learning field, to improve their interpretability and performance on visual data. In this thesis, we optimize a recent biologically inspired deep learning architecture designed for analysis of population data recorded from mammalian primary visual cortex when presented with natural images as stimuli. We reimplement this prior modeling in existing neuroscience focused deep learning framework NDN3 and assess it in terms of stability and sensitivity to hyperparameters and architecture fine-tuning. We proceed to extend the model with components of various DNN models, analysing novel combinations and techniques from classical computer vision deep learning, comparing their effectiveness against the bio-inspired components. We were able to identify modifications that greatly increase the stability of the model while securing moderate improvement in overall performance. Furthermore, we document the importance of small hyperparameters adjustments versus...

Nesta tese estudamos o fluxo dirigido, a primeira componente na expansão de Fourier da distribuição azimutal das partículas emitidas. Diferente do fluxo elíptico, que é muito bem estudado e descrito na literatura, a física que gera este observável ainda não é muito conhecida. Mostramos que este observável é altamente sensível a condições iniciais, comparando vários resultados numéricos para diferentes condições iniciais, entre analíticas e numéricas. Propomos que o fluxo dirigido também é sensível à aceleração longitudinal e formulamos um modelo analítico baseado nesta hipótese. Este modelo sugere quais ingredientes são relevantes para o fluxo dirigido. Ele é confrontado com sucesso com cálculos numéricos, resultados experimentais e cálculos que não incluem a aceleração longitudinal. / In this thesis we study the directed flow, the first component in the Fourier\'s expansion of the azimuthal distribution of emitted particles. Unlike the elliptic flow, which is well studied and described in the literature, the physics that generates this observable is not yet well described. We show that this observable is highly sensible to the initial conditions, comparing several numeric results with different initial conditions, between analytic and numeric ones. We propose that the directed flow is also sensitive to the longitudinal acceleration and we formulate an analytic model based in this hypothesis. This model suggests which ingredients are relevant to the directed flow. It is confronted with success against numeric calculus, experimental results, and calculus that do not include the longitudinal acceleration.

Dans cette thèse, nous développons une approche supervisée de reconnaissance d’objets basée sur l’utilisation de nouveaux descripteurs d’images globaux inspirés du modèle du cortex visuel humain primaire V1 en tant que groupe de roto-translations semi-discrètes SE (2,N)=R² x ZN produit semi-direct entre R² et ZN. La méthode proposée est basée sur des descripteurs de Fourier généralisés et rotationnels définis sur le groupe SE (2,N), qui sont invariants aux transformations géométriques (translations, et rotations). De plus, nous montrons que ces descripteur de Fourier sont faiblement complets, dans le sens qu’ils permettent de discriminer sur un ensemble ouvert et dense L² (SE(2,N)) de fonctions à support compact, donc distinguer entre des images réelles. Ces descripteurs sont ensuite utilisés pour alimenter un classifieur de type SVM dans le cadre de la reconnaissance d’objets. Nous avons mené une séries d’expérimentations dans le but d’évaluer notre méthode sur les bases de visages RL, CVL et ORL et sur la base d’images d’objets variés COIL-100, et de comparer ses performances à celles des méthodes basées sur des descripteurs globaux et locaux. Les résultats obtenus ont montré que notre approche est en mesure de concurrencer de nombreuses techniques de reconnaissance d’objets existantes et de surpasser de nombreuse autres. Ces résultats ont également montré que notre méthode est robuste aux bruits. Enfin, nous avons employé la technique proposée pour reconnaître des navires dans un contexte de surveillance maritime. / In this thesis, we develop a supervised object recognition method using new global image descriptors inspired by the model of the human primary visual cortex V1. Mathematically speaking, the latter is modeled as the semi-discrete roto-translation group SE (2,N)=R² x ZN semi-direct product between R² and ZN. Therefore, our technique is based on generalized and rotational Fourier descriptors defined in SE (2,N) , and which are invariant to natural geometric transformations (translations, and rotations). Furthermore, we show that such Fourier descriptors are weakly complete, in the sense that they allow to distinguish over an open and dense set of compactly supported functions in L² (SE(2,N)) , hence between real-world images. These descriptors are later used in order to feed a Support Vector Machine (SVM) classifier for object recognition purposes. We have conducted a series of experiments aiming both at evaluating and comparing the performances of our method against existing both local - and global - descriptor based state of the art techniques, using the RL, the CVL, and the ORL face databases, and the COIL-100 image database (containing various types of objects). The obtained results have demonstrated that our approach was able to compete with many existing state of the art object recognition techniques, and to outperform many others. These results have also shown that our method is robust to noise. Finally, we have applied the proposed method on vessels recognition in the framework of maritime surveillance.

Nesta tese estudamos o fluxo dirigido, a primeira componente na expansão de Fourier da distribuição azimutal das partículas emitidas. Diferente do fluxo elíptico, que é muito bem estudado e descrito na literatura, a física que gera este observável ainda não é muito conhecida. Mostramos que este observável é altamente sensível a condições iniciais, comparando vários resultados numéricos para diferentes condições iniciais, entre analíticas e numéricas. Propomos que o fluxo dirigido também é sensível à aceleração longitudinal e formulamos um modelo analítico baseado nesta hipótese. Este modelo sugere quais ingredientes são relevantes para o fluxo dirigido. Ele é confrontado com sucesso com cálculos numéricos, resultados experimentais e cálculos que não incluem a aceleração longitudinal. / In this thesis we study the directed flow, the first component in the Fourier\'s expansion of the azimuthal distribution of emitted particles. Unlike the elliptic flow, which is well studied and described in the literature, the physics that generates this observable is not yet well described. We show that this observable is highly sensible to the initial conditions, comparing several numeric results with different initial conditions, between analytic and numeric ones. We propose that the directed flow is also sensitive to the longitudinal acceleration and we formulate an analytic model based in this hypothesis. This model suggests which ingredients are relevant to the directed flow. It is confronted with success against numeric calculus, experimental results, and calculus that do not include the longitudinal acceleration.

This thesis investigated the processing of emotional stimuli by the visual system, and how the processing of emotions interacts with visual awareness. Emotions have been given ‘special’ status by some previous research, with evidence that the processing of emotions may be relatively independent of striate cortex, and less affected by disruption to awareness than processing of emotionally neutral images. Yet the extent to which emotions are ‘special’ remains questionable. This thesis focused on the processing of emotional stimuli when activity in V1 was disrupted using transcranial magnetic stimulation (TMS), and whether emotional properties of stimuli can be reliably discriminated, or affect subsequent responses, when visibility is low. Two of the experiments reported in this thesis disrupted activity in V1 using TMS, Experiment 1 with single pulses in an online design, and Experiment 2 with theta burst stimulation in an offline design. Experiment 1 found that a single pulse of TMS 70-130 ms following a presentation of a body posture image disrupted processing of neutral but not emotional postures in an area of the visual field that corresponded to the disruption. Experiment 2 did not find any convincing evidence of disruption to processing of neutral or emotional faces. From Experiment 1 it would appear that emotional body posture images were relatively unaffected by TMS, and appeared to be robust to disruption to V1. Experiment 2 did not add to this as there was no evidence of disruption in any condition. Experiments 3 and 4 used visual masking to disrupt awareness of emotional and neutral faces. Both experiments used a varying interval between the face and the mask stimuli to systematically vary the visibility of the faces. Overall, the shortest SOA produced the lowest level of visibility, and this level of visibility was arguably outside awareness. In Experiment 3, participants’ ability to discriminate properties of emotional faces under low visibility conditions was greater than their ability to discriminate the orientation of the face. This was despite the orientation discrimination being much easier at higher levels of visibility. Experiment 4 used a gender discrimination task, with emotion providing a redundant cue to the decision (present half of the time). Despite showing a strong linear masking function for the neutral faces, there was no evidence of any emotion advantage. Overall, Experiment 3 gave some evidence of an emotion advantage under low visibility conditions, but this effect was fairly small and not replicated in Experiment 4. Finally, Experiments 5-8 used low visibility emotional faces to prime responses to subsequent emotional faces (Experiments 5 and 6) or words (Experiments 7 and 8). In Experiments 5, 7 and 8 there was some evidence of emotional priming effects, although these effects varied considerably across the different designs used. There was evidence for meaningful processing of the emotional prime faces, but this processing only led to small and variable effects on subsequent responses. In summary, this thesis found some evidence that the processing of emotional stimuli was relatively robust to disruption in V1 with TMS. Attempts to find evidence for robust processing of emotional stimuli when disrupted with backwards masking was less successful, with at best mixed results from discrimination tasks and priming experiments. Whether emotional stimuli are processed by a separate route(s) in the brain is still very much open to debate, but the findings of this thesis offers small and inconsistent evidence for a brain network for processing emotions that is relatively independent of V1 and visual awareness. The network and nature of brain structures involved in the processing of subliminal and low visibility processing of emotions remains somewhat elusive.

Tese (Doutorado) - Naval Postgraduate Scholl, Monterey, CA, 2017. / Devido a falta de auxílios na navegação subaquática, um sistema capaz de ajudar um veículo submersível a manter a precisão da posição, permitindo que ele permaneça subaquático por períodos mais longos, é extremamente desejável. A precisão e confiabilidade / Made available in DSpace on 2018-02-14T17:42:12Z (GMT). No. of bitstreams: 0 Previous issue date: 2017. Added 1 bitstream(s) on 2019-02-13T19:12:12Z : No. of bitstreams: 1 000015dd.pdf: 3906657 bytes, checksum: f7516e210feff674af44c4ed61eae57f (MD5)

No description available.

No description available.