Submitted by Luciana Sebin (lusebin@ufscar.br) on 2016-10-06T13:29:02Z No. of bitstreams: 1 DissBCGS.pdf: 7250050 bytes, checksum: df4b2203e5e586a2cba2f75ff4af7f2f (MD5) / Approved for entry into archive by Marina Freitas (marinapf@ufscar.br) on 2016-10-13T20:24:11Z (GMT) No. of bitstreams: 1 DissBCGS.pdf: 7250050 bytes, checksum: df4b2203e5e586a2cba2f75ff4af7f2f (MD5) / Approved for entry into archive by Marina Freitas (marinapf@ufscar.br) on 2016-10-13T20:24:21Z (GMT) No. of bitstreams: 1 DissBCGS.pdf: 7250050 bytes, checksum: df4b2203e5e586a2cba2f75ff4af7f2f (MD5) / Made available in DSpace on 2016-10-13T20:24:30Z (GMT). No. of bitstreams: 1 DissBCGS.pdf: 7250050 bytes, checksum: df4b2203e5e586a2cba2f75ff4af7f2f (MD5) Previous issue date: 2016-02-19 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / Over the last few years, a large number of researchers have directed their efforts and interests for the in vivo study of the cellular and molecular mechanisms of leukocyte-endothelial interactions in the microcirculation of many tissues under different inflammatory conditions. The main goal of these studies is to develop more effective therapeutic strategies for the treatment of inflammatory and autoimmune diseases. Nowadays, analysis of the leukocyte-endothelial interactions in small animals is performed by visual assessment from an intravital microscopy image sequences. Besides being time consuming, this procedure may cause visual fatigue of the observer and, therefore, generate false statistics. In this context, this work aims to study and develop computational techniques for the automatic detection and tracking of leukocytes in intravital video microscopy. For that, results from frame to frame processing (2D – spatial analysis) will be combined with those from the three-dimensional analysis (3D=2D+t – spatio-temporal analysis) of the volume formed by stacking the video frames. The main technique addressed for both processings is based on the analysis of the eigenvalues of the local Hessian matrix. While the 2D image processing aims the leukocyte detection without worrying about their tracking, 2D+t processing is intended to assist on the dynamic analysis of cell movement (tracking), being able to predict cell movements in cases of occlusion, for example. In this work we used intravital video microscopy obtained from a study of Multiple Sclerosis in mice. Noise reduction and registration techniques comprise the preprocessing step. In addition, techniques for the analysis and definition of cellular pathways comprise the post processing step. Results of 2D and 2D+t processing steps, compared with conventional visual analysis, have shown the effectiveness of the proposed approach. / Nos últimos anos, um grande número de pesquisadores tem voltado seus esforços e interesses para o estudo in vivo dos mecanismos celulares e moleculares da interação leucócitoendotélio na microcirculação de vários tecidos e em várias condições inflamatórias. O principal objetivo desses estudos é desenvolver estratégias terapêuticas mais eficazes para o tratamento de doenças inflamatórias e autoimunes. Atualmente, a análise de interações leucócito-endotélio em pequenos animais é realizada de maneira visual a partir de uma sequência de imagens de microscopia intravital. Além de ser demorado, esse procedimento pode levar à fadiga visual do observador e, portanto, gerar falsas estatísticas. Nesse contexto, este trabalho de pesquisa tem como objetivo estudar e desenvolver técnicas computacionais para a detecção e rastreamento automáticos de leucócitos em vídeos de microscopia intravital. Para isso, resultados do processamento quadro a quadro do vídeo (2D – análise espacial) serão combinados com os resultados da análise tridimensional (3D=2D+t – análise espaço-temporal) do volume formado pelo empilhamento dos quadros do vídeo. A principal técnica abordada para ambos os processamentos é baseada na análise local dos autovalores da matriz Hessiana. Enquanto o processamento de imagens 2D tem como objetivo a detecção de leucócitos sem se preocupar com seu rastreamento, o processamento 2D+t pretende auxiliar na análise dinâmica de movimentação das células (rastreamento), sendo capaz de prever movimentos celulares em casos de oclusão, por exemplo. Neste trabalho foram utilizados vídeos de microscopia intravital obtidos a partir de um estudo da Esclerose Múltipla realizado com camundongos. Técnicas de redução de ruído e estabilização do movimento das sequências de imagens compõem a etapa de pré-processamento, assim como técnicas para a definição e análise dos caminhos celulares compõem a etapa de pós-processamento. Resultados das etapas de processamento 2D e 2D+t, comparados com a convencional análise visual, demonstraram a eficácia da abordagem proposta. / FAPESP: 2013/26171-6

This master‘s thesis deals with the Deep Learning methods for image recognition tasks from the first methods to the modern ones. The main focus is on convolutional neural nets based models for classification, detection and image segmentation. These methods are used for practical implemetation – counting passing cars on video from traffic camera. After several test of available models, the YOLOv2 architecture was chosen and retrained on own dataset. The application also includes the addition of the SORT tracking algorithm.

The goal of this thesis is to design and create an automatic system for video anonymization. This system makes use of various object detectors on an image to ensure functionality, as well as active tracking of objects detected in this manner. Adjustments are later applied to these detected objects which ensure sufficient level of anonymization. The main asset of this system is speeding up the anonymization process of videos that can be published after.

The aim of this diploma thesis is to obtain information on the motion of people in a scene from the record of the stationary camera. The procedure to detect exceptional events in the scene was designed. Exceptional events can be fast-moving persons, or persons moving in di erent places than everyone else in the scene. To trace the motion of persons, two algorithms were applied and tested - Optical flow and CAMSHIFT. The analysis of the resulting motions is performed by monitoring the progress of motion, and its comparison with the other motions in the scene. The analysis result is represented by detected exceptional motions that can be found in the video. The areas where the motion occurs in the scene, and where the motion is the most common are also described together with the motion direction analysis. The exceptional motion parts extracted from the video represent the main result of the work.

Real-time lane detection or localization is a crucial problem in modern Advanced Driver Assistance Systems (ADAS), especially in Automated Driving System. This thesis estimates the possibility to implement a lane detection system in the available low-power embedded hardware. Various state-of-the-art Lane Detection algorithms are assessed based on a number of proposed criteria. From the result of the evaluation, three different algorithms are constructed and implemented in the hardware using OpenCV library. The lane detection stage is done with different methods: using Hough Transform for line detection or randomly sampling hypotheses which are straight lines or cubic splines over the pre-processed binary image. Weights of the hypotheses are then calculated based on their positions in the image. The hypothesis which has highest weight or best position will be chosen to represent lane marking. To increase the performance of the system, tracking stage is introduced with the help of Particle Filter or Kalman Filter. The systems are then tested with several different datasets to evaluate the speed, performance and ability to work in real-time. In addition, the system interfaces with CAN bus using a CAN interface, so that the output data can be sent as messages via the CAN bus to other systems.

L’objectif de ces travaux porte sur la navigation de robots autonomes sur de grandes distances dans des environnements extérieurs dynamiques, plus précisément sur le développement et l’évaluation de fonctions avancées de perception, embarquées sur des véhicules se déplaçant en convoi sur un itinéraire inconnu a priori, dans un environnement urbain ou naturel. Nous avons abordé trois problématiques : d’abord nous avons exploité plusieurs méthodes de l’état de l’art, pour qu’un véhicule A, équipé d’un capteur stéréoscopique, apprenne à la fois une trajectoire et un modèle de l’environnement supposé d’abord statique. Puis nous avons proposé deux modes pour l’exécution de cette trajectoire par un véhicule B équipé d’une simple caméra : soit un mode différé, dans lequel B charge toute la trajectoire apprise par A, puis l’exécute seul, soit un mode convoi, dans lequel B suit A, qui lui envoie par une communication HF, les tronçons de la trajectoire au fur et à mesure qu’ils sont appris. Enfin nous avons considéré le cas des environnements évolutifs et dynamiques, en traitant de la détection d’événements depuis les images acquises depuis un véhicule mobile: détection des changements (disparition ou apparition d’objets statiques, typiquement des véhicules garés dans un milieu urbain), ou de la détection d’objets mobiles (autres véhicules ou piétons) / The global objective of these works concerns the navigation of autonomous robots on long routes in outdoor dynamic environments, more precisely on the development and the evaluation of advanced perception functions, embedded on vehicles moving in a convoy formation, on an a priori unknown route in urban or natural environments. Three issues are tackled: first several methods from the State of the Art have been integrated in order to cope with the visual mapping and the trajectory learning problems for a vehicle A equipped with a stereovision sensor, moving in a large-scale environment, assumed static. Then it is proposed two modes for the execution of this trajectory by a vehicle B equipped by a single camera: either a delayed mode, in which B loads initially all representations learnt by A, and executes alone the recorded trajectory, or a convoy mode, in which B follows A, which sends him by a communication link, the trajectory sections as soon as they are learnt. Finally, it has been considered changing and dynamic environments, dealing with the detection of events from images acquired on a dynamic vehicle: detection of changes (disappearances or appearances of static objects, typically cars parked in a urban environment), or detection of mobile objects (pedestrians or other vehicles)

No description available.

Augmented reality has been on the rise for some time now and begun making its way onto the mobile market for both IOS and Android. In 2017 Apple released ARKit for IOS which is a software development kit for developing augmented reality applications. To counter this, Google released their own variant called ARCore on the 1st of march 2018. ARCore is also a software development kit for developing augmented reality applications but made for the Android, Unity and Unreal platforms instead. Since ARCore is released recently it is still unknown what particular limitations may exist for it. The purpose of this paper is give an indication to companies and developers about ARCore’s potential limitations. The goal with this paper and work is to map how well ARCore works during different circumstances, and in particular, how its feature detection works and behaves. A quantitative research was done with the usage of the case study method. Various tests were performed with a modified test-application supplied by Google. The tests included testing how ARCore’s feature detection, the process that analyzes the environment presented to the application. This which enables the user of an application to place a virtual object on the physical environment. The tests were done to see how ARCore works during different light levels, different types of surfaces, different angles and the difference between having the device stationary or moving. From the testing that were done some conclusions could be drawn about the light levels, surfaces and differences between a moving and stationary device. More research and testing following these principles need to be done to draw even more conclusions of the system and its limitations. How these should be done is presented and discussed. / Forstarkt verklighet (augmented reality) har stigit under en tid och börjat ta sig in på mobilmarknaden for både IOS och Android. År 2017 släppte Apple ARKit för IOS vilket är en utvecklingsplattform för att utveckla applikationer inom förstärkt verklighet. Som svar på detta så slappte Google sin egna utvecklingsplattform vid namn ARCore, som släpptes den 1 mars 2018. ARCore är också en utvecklingsplattform för utvecklandet av applikationer inom förstarkt verklighet men istället inom Android, Unity och Unreal. Sedan ARCore släpptes nyligen är det fortfarande okant vilka särskilda begränsningar som kan finnas för det. Syftet med denna avhandling är att ge företag och utvecklare en indikation om ARCores potentiella begränsningar. Målet med denna avhandling och arbete är att kartlägga hur väl ARCore fungerar under olika omstandigheter, och i synnerhet hur dess struktursdetektor fungerar och beter sig. En kvantitativ forskning gjordes med användning av fallstudie metoden. Olika tester utfördes med en modifierad test-applikation från Google. Testerna inkluderade testning av hur ARCores struktursdetektor, processen som analyserar miljön runt om sig, fungerar. Denna teknik möjliggor att användaren av en applikation kan placera ett virtuellt objekt på den fysiska miljön. Testen innebar att se hur ARCore arbetar under olika ljusnivåer, olika typer av ytor, olika vinklar och skillnaden mellan att ha enheten stationär eller rör på sig. Från testningen som gjordes kan man dra några slutsatser om ljusnivåer, ytor och skillnader mellan en rörlig och stationar enhet. Mer forskning och testning enligt dessa principer måste göras för att dra ännu mer slutsatser av systemet och dess begränsningar. Hur dessa ska göras presenteras och diskuteras.

Les travaux présentés dans ce mémoire s’inscrivent dans le cadre duprojet PANsafer (Vers un Passage A Niveau plus sûr), lauréat de l’appel ANR-VTT2008. Ce projet est labellisé par les deux pôles de compétitivité i-Trans et Véhiculedu Futur. Le travail de la thèse est mené conjointement par le laboratoire IRTESSETde l’UTBM et le laboratoire LEOST de l’IFSTTAR.L’objectif de cette thèse est de développer un système de perception permettantl’interprétation de scénarios dans l’environnement d’un passage à niveau. Il s’agitd’évaluer des situations potentiellement dangereuses par l’analyse spatio-temporelledes objets présents autour du passage à niveau.Pour atteindre cet objectif, le travail est décomposé en trois étapes principales. Lapremière étape est consacrée à la mise en place d’une architecture spatiale des capteursvidéo permettant de couvrir de manière optimale l’environnement du passageà niveau. Cette étape est mise en oeuvre dans le cadre du développement d’unsimulateur d’aide à la sécurité aux passages à niveau en utilisant un système deperception multi-vues. Dans ce cadre, nous avons proposé une méthode d’optimisationpermettant de déterminer automatiquement la position et l’orientation descaméras par rapport à l’environnement à percevoir.La deuxième étape consisteà développer une méthode robuste de suivi d’objets enmouvement à partir d’une séquence d’images. Dans un premier temps, nous avonsproposé une technique permettant la détection et la séparation des objets. Le processusde suivi est ensuite mis en oeuvre par le calcul et la rectification du flotoptique grâce respectivement à un modèle gaussien et un modèle de filtre de Kalman.La dernière étape est destinée à l’analyse des trajectoires 2D reconstruites parl’étape précédente pour l’interprétation de scénarios. Cette analyse commence parune modélisation markovienne des trajectoires 2D. Un système de décision à basede théorie de l’évidence est ensuite proposé pour l’évaluation de scénarios, aprèsavoir modélisé les sources de danger.L’approche proposée a été testée et évaluée avec des données issues de campagnesexpérimentales effectuées sur site réel d’un passage à niveau mis à disposition parRFF. / The main objective of this thesis is to develop a system for monitoringthe close environment of a level crossing. It aims to develop a perception systemallowing the detection and the evaluation of dangerous situations around a levelcrossing.To achieve this goal, the overall problem of this work has been broken down intothree main stages. In the first stage, we propose a method for optimizing automaticallythe location of video sensors in order to cover optimally a level crossingenvironment. This stage addresses the problem of cameras positioning and orientationin order to view optimally monitored scenes.The second stage aims to implement a method for objects tracking within a surveillancezone. It consists first on developing robust algorithms for detecting and separatingmoving objects around level crossing. The second part of this stage consistsin performing object tracking using a Gaussian propagation optical flow based modeland Kalman filtering.On the basis of the previous steps, the last stage is concerned to present a newmodel to evaluate and recognize potential dangerous situations in a level crossingenvironment. This danger evaluation method is built using Hidden Markov Modeland credibility model.Finally, synthetics and real data are used to test the effectiveness and the robustnessof the proposed algorithms and the whole approach by considering various scenarioswithin several situations.This work is developed within the framework of PANsafer project (Towards a saferlevel crossing), supported by the ANR-VTT program (2008) of the French NationalAgency of Research. This project is also labelled by Pôles de compétitivité "i-Trans"and "Véhicule du Futur". All the work, presented in this thesis, has been conductedjointly within IRTES-SET laboratory from UTBM and LEOST laboratory fromIFSTTAR.

近年來科技、網路以及儲存媒介的發達，產生的資料量呈現爆炸性的成長，也宣告了巨量資料時代的來臨。擁有巨量資料代表了不必再依靠傳統抽樣的方式來蒐集資料，分析數據也不再有資料收集不足以致於無法代表母題的限制。突破傳統的限制後，巨量資料的精隨在於如何從中找出有價值的資訊。 以擁有大量輿論和人際互動資訊的社群網站為例，就有相關學者研究其情緒與股價具有正相關性，本研究也試著利用同樣具有巨量資料特性的網路新聞，抓取中央新聞社2013年7月至2014年5月之經濟類新聞共計30,879篇，結合新聞主題偵測與追蹤技術及情感分析，利用新聞事件相似的概念，透過連結匯聚成網絡並且分析新聞的情緒和股價指數的關係。 研究結果顯示，新聞事件間可以連結成一特定新聞主題，且能在龐大的網絡中找出不同的新聞主題，並透過新聞主題之連結產生新聞主題脈絡。對此提供一種新的方式來迅速了解巨量新聞內容，也能有效的回溯新聞主題及新聞事件。 在新聞情緒和股價指數方面，研究發現新聞情緒影響了股價指數之波動，其相關係數達到0.733562；且藉由情緒與心理線及買賣意願指標之比較，顯示新聞的情緒具有一定的程度能夠成為股價判斷之參考依據。 / In recent years, the technology, network, and storage media developed, the amount of generated data with the explosive growth, and also declared the new era of big data. Having big data let us no longer rely on the traditional sample ways to collect data, and no longer have the issue that could not represent the population which caused by the inadequate data collection. Once we break the limitations, the main spirit of big data is how to find out the valuable information in big data. For example, the social network sites (SNS) have a lot of public opinions and interpersonal information, and scholars have founded that the emotions in SNS have a positive correlation with stock prices. Therefore, the thesis tried to focus on the news which have the same characteristic of big data, using the web crawl to catch total of 30,879 economics news articles form the Central News Agency, furthermore, took the “Topic Detection & Tracking” and “Sentiment Analysis” technology on these articles. Finally, based on the concept of the similarity between news articles, through the links converging networks and analyze the relevant between news sentiment and stock prices. The results shows that news events can be linked to specific news topics, identify different news topics in a large network, and form the news topic context by linked news topics together. The thesis provides a new way to quickly understand the huge amount of news, and backtracking news topics and news event with effective. In the aspect of news sentiment and stock prices, the results shows that the news sentiments impact the fluctuations of stock prices, and the correlation coefficient is 0.733562. By comparing the emotion with psychological lines & trading willingness indicators, the emotion is better than the two indicators in the stock prices determination.