The packing and coding of digital video is a part of science where much innovation has taken place during the last few decades. The MPEG standards of video encoding are some of the most well-known and used video coding standards today. Since MPEG defines exact requirements for the decoder, but not the encoder, encoders can be made in many different ways and levels of complexity, as long as they produce legal MPEG streams that can be viewed on any MPEG-conformant decoder. This thesis will examine the possible performance of MPEG, in particular MPEG-1/MPEG-2 full TV resolution (720*576), for coding video at bitrates significantly lower than the 2-15 Mb/s MPEG-2 originally was designed for. For this purpose, encoding methods previously proposed by various researchers are presented. Furthermore a few new algorithms, which can be used for MPEG coding in general, but was constructed with a low-bitrate encoder in mind, were developed. Finally objective video quality benchmarks were conducted and the results evaluated.

We propose a novel approach for segmentation of human objects, including face and body, in image sequences. In modern video coding techniques, e.g., MPEG-4 and MPEG-7, human objects are usually the main focus for multimedia applications. We combine temporal and spatial information and employ a neuro-fuzzy mechanism to extract human objects. A fuzzy self-clustering technique is used to divide the video frame into a set of segments. The existence of a face within a candidate face region is ensured by searching for possible constellations of eye-mouth triangles and verifying each eye-mouth combination with the predefined template. Then rough foreground and background are formed based on a combination of multiple criteria. Finally, human objects in the base frame and the remaining frames of the video stream are precisely located by a fuzzy neural network which is trained by a SVD-based hybrid learning algorithm. Through experiments, we compare our system with two other approaches, and the results have shown that our system can detect face locations and extract human objects more accurately.

Efficient and effective vision systems are proposed in this work for object detection for ground&aerial robots venturing into unknown environments with minimum vision aids, i.e. a single camera. The first problem attempted is that of object search and identification in a situation similar to a disaster site. Based on image analysis, typical pixel-based characteristics of a visual marker have been established to search for, using a block based search algorithm, along with a noise and interference filter. The proposed algorithm has been successfully utilized for the International Aerial Robotics competition 2009. The second problem deals with object detection for collision avoidance in 3D environments. It has been shown that a 3D model of the scene can be generated from 2D image information from a single camera flying through a very small arc of lateral flight around the object, without the need of capturing images from all sides. The forward flight simulations show that the depth extracted from forward motion is usable for large part of the image. After analyzing various constraints associated with this and other existing approaches, Motion Estimation has been proposed. Implementation of motion estimation on videos from onboard cameras resulted in various undesirable and noisy vectors. An in depth analysis of such vectors is presented and solutions are proposed and implemented, demonstrating desirable motion estimation for collision avoidance task.

In dieser Arbeit wird das Real-Time Streaming Protocol (RTSP) spwie das Real-Time Transport Protocol (RTP) analysiert. Die Syntax und Semantik von Präsentations-Beschreibungen im Format des Session Description Protocol (SDP) werden vorgestellt. Es wird eine kurze Einführung in das Format von MPEG-1-Videosequenzen gegeben. In der Analysephase wird der aktuelle Stand des Inline-MPEG-1-Players und des Multicast MPEG-Servers untersucht. Ausgehend davon werden die Ziele der Implementierung eines neuen Java-Applets sowie eines RTSP-Servers festgelegt und beschrieben, in welchen Schritten diese erreicht werden sollen. In der Implementierungsphase werden die Erstellung eines RTSP- und eines RTP-Klienten in Java beschrieben. Es wird dargelegt, wie unter Verwendung dieser Klienten auf Basis des Inline-MPEG-1-Players ein Java-Applet erzeugt wurde, daß MPEG-1-Videosequenzen von einem Medienserver empfangen und abspielen kann. Es wird der Entwurf eines RTSP-Medienservers und dessen Implementierung beschrieben. Mit Abschluß dieser Arbeit steht ein einsatzbereites System aus RTSP-Server und RTSP-Klient bereit, daß ohne weitere Klientsoftware auf einer Weboberfläche eingesetzt werden kann.

Για την προστασία της χρήσης ψηφιακών αντικειμένων έχουν αναπτυχθεί διάφορα συστήματα Ψηφιακής Διαχείρισης Δικαιωμάτων (Digital Rights Management). Υπάρχουν πολλές τεχνολογικές λύσεις που μειονεκτούν όμως όσον αφορά τη μεταξύ τους διαλειτουργικότητα. Το Motion Pictures Expert Group με το τελευταίο του πρότυπο MPEG-21 δημιουργεί ένα πλαίσιο που διευκολύνει την διαλειτουργικότητα των εφαρμογών αυτών. Στο μέρος 5 του προτύπου αναπτύσει ένα πρότυπο για μία γλώσσα έκφρασης δικαιωμάτων (Rights Expression Language). Αντικείμενο της εργασίας είναι ένα πρόγραμμα υπολογιστή που υλοποιεί τον πυρήνα αυτού του προτύπου και επιλεγμένα χαρακτηριστικά του. Κατά τη λειτουργία του ικανοποιεί την αίτηση του χρήστη για την παραγωγή μίας MPEG-21 άδειας χρήσης. Η έγκριση και παραγωγή της άδειας εξαρτάται από τους κανόνες λειτουργίας (business rules) που έχουν τροφοδοτηθεί στην εφαρμογή. / For the protection of digital items eist many diverse DRM technologies. MPEG has ndertaken the effort to provide interoperability for such systems as part of mpeg-21 effort. The object of this work is a piece of computer software that realizes the authorization algorithm of mpeg21 part5 and issues a mpeg21 license. It also supports the variables mechanism defined in mpeg21 part5

In the last few years 3D Virtual Worlds (3DVWs) became a reality. Initially considered as a new mean for social communication, triggered by the development of software and hardware technology, 3DVWs are exposing now different functionalities, experiences and acquaintances. Therefore, they achieved their popularity very fast, indicated by the number and the progression of active users. Being the representation of the user, the avatar is one of the most significant and most complex assets of a Virtual World. A short analysis of a VW content allows one to observe that from the point of view of the storage/transmission the most significant amount is represented by the VW assets. Within the set of assets, the avatars are the most complex structures, consisting of different components: geometry, images, animations, structures, etc. The overall objective of developing tools and methods for a large deployment of VW are translated into three specific ones: To propose a compression framework to enable efficient, compact transfer of avatars, and general 3D graphics assets. Specifically for avatars, the framework should be independent from the representation formalism. To propose an optimized solution making the avatars accessible on weak terminals such as mobile phones. To define a metadata model allowing avatars interoperability between different VWs. The three objectives are addressed in this thesis and for each we propose original contributions.

Technologijos greitai keičiasi, kasdien atsiranda mokslo ir technikos naujovių, kurios daugiau ar mažiau įtakoja mūsų kasdieninį gyvenimą. Skaitmeninės technologijos atnešė daug naujovių ir galima numanyti, kad jų bus dar daugiau. Skaitmenizavimo procesas tęsiasi ir apima dar likusias analogines erdves. Ypatingai šis procesas svarbus komunikacijų srityje Darbe stengtasi išsiaiškinti ir ištirti DVI veikimo principus bei suprojektuoti šios sąsajos modifikaciją, besiremiančia suspaustų vaizdų perdavimo technologija. Darbe susiduriama su vaizdo duomenų formavimo, perdavimo ir atvaizdavimo mechanizmais. Tuo pat panagrinėjant įvairius nestandartinius šio mechanizmo atvejus, kurie galbūt dabar nėra komerciškai efektyvūs, bet galėtų rasti labai specifinį panaudojimą ir būtent jame taptų nepakeičiamai naudingi. Nuspręsta imtis skaitmeninės DVI sąsajos (angl. digital visual interface). Atliekamų funkcijų atskyrimui naudojama blokinė dekompozicija, apžvelgiamos aparatūrinės ir programinės užsibrėžtos sistemos kūrimo priemonės. Kadangi DVI sąsaja jungiamas kompiuteris ir skaitmeninis monitorius paprastai yra vienas nuo kito netoli, nuspręsta pabandyti kaip pavyktų šį atstumą pailginti neapibrėžtai daug, kitaip tariant vaizdą perduoti naudojant įprastą ethernet tipo laidą, duomenų formatą pakeičiant į paketinį - IP (interneto protokolą) bei su kokiomis pagrindinėmis problemomis susiduriama tai darant. Darbe nagrinėjama ir iškeltos funkcijos ar jos dalies perkėlimo į lustą galimybė. / The study tries to clear out how does the digital video data transfer system works. The scope is about DVI interface and similar systems. Also the design of compressed video data transfer mechanisms. The things touched are very different and not even very popular in deed but in some cases they can become a very useful decision for some non standard applications. The study mainly touches digital visual interface – DVI. In case that DVI interface connected computer and monitor must be really near each other it was decided to enlong this distance. In other words – to transfer video data through Ethernet cable and changing and compressing DVI video data format into packets. All study stands for it and tries to clarify the main problems and difficulties doing that.

A esteganografia em vídeos digitais possibilita o ocultamento de um grande volume de informações quando comparada a técnicas em imagens. Contudo, esta tarefa não é trivial quando aplicada a vídeos comprimidos, pois a inserção de informações ocultas pode adicionar ruído dificultando a perfeita recuperação das mesmas durante a decodificação. Este trabalho apresenta uma técnica para esteganografia em vídeos comprimidos, denominada MP4Stego, a qual explora estruturas e tecnologias do padrão de vídeo MPEG-4 de modo a proceder recuperação de informações sem perda e apresentar maior capacidade de inserção de dados ocultos. Entre os benefícios da técnica estão: A capacidade de ocultamento de grande volume de dados; a capacidade de reprodução do vídeo contendo dados ocultos em players não especializados (ajudando a dar a ilusão de que se trata de um vídeo comum), sua imunidade, até o momento, as técnicas de esteganálise / Digital video steganography makes possible to hide higher volumes of information than the broadly used image steganography techniques. However, apply steganography to compressed digital videos is not a trivial task because the information hiding process may add noise to the original data. In this way, during the video decoding, the recovered hidden information is not an exact copy of the original data. This work presents a new steganographic technique, called MP4Stego, which explores new video structures of the MPEG-4 standard in order to recover hidden data without loss and in order to provide higher capacity of hidden data insertion. Some benefits of the MP4Stego technique are: to hide high volumes of data; to playback steganographyed videos by the means of non specialized players (this helps users to think that it is a regular video); so far, its immunity to stegoanalysis techniques

Submitted by Renata Lopes (renatasil82@gmail.com) on 2017-05-30T15:08:44Z No. of bitstreams: 1 rodrigodamascenomarangon.pdf: 825004 bytes, checksum: f6fba786f13817ab4b7d8d41a6cf4d04 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2017-05-30T15:36:45Z (GMT) No. of bitstreams: 1 rodrigodamascenomarangon.pdf: 825004 bytes, checksum: f6fba786f13817ab4b7d8d41a6cf4d04 (MD5) / Made available in DSpace on 2017-05-30T15:36:45Z (GMT). No. of bitstreams: 1 rodrigodamascenomarangon.pdf: 825004 bytes, checksum: f6fba786f13817ab4b7d8d41a6cf4d04 (MD5) Previous issue date: 2016-08-26 / O consumo de multimídia pela web tem crescido fortemente ao longo dos últimos anos. Com isso surge a preocupação do consumo de banda, que é significativamente reforçada pelo aumento da utilização da computação móvel. É fato conhecido que o consumo de banda para a multimídia ultrapassou o consumo nas redes P2P, e tende ao crescimento nos próximos anos. Nesse cenário, as tecnologias de streaming multimídia via HTTP vieram como parte da solução do problema de maior demanda de tráfego de multimídia. O MPEG-DASH é uma dessas tecnologias, e atualmente é utilizada em grande parte por grandes corporações que distribuem multimídia, como por exemplo Google e Netflix. No entanto, o processo da criação das mídias no padrão MPEG-DASH é um processo trabalhoso e de pouca documentação. O presente trabalho apresenta técnicas para transcodificação, segmentação e reprodução de mídias no padrão MPEG-DASH. As mídias convertidas são parte do material didático produzido pelo Centro de Educação a Distância da Universidade Federal de Juiz de Fora. O estudo foca no processo técnico de padronização das mídias para MPEG-DASH, dá foco aos problemas encontrados, além de analisar aspectos da reprodução das mídias. A dissertação traz também um estudo de caso, e análise de aspectos relevantes relacionados ao ambiente computacional da pesquisa. Foram testadas diversas técnicas e ferramentas, e os problemas, soluções e sugestões de uso são relatados ao longo do trabalho. Os resultados mostram que os resultados, apesar de problemas pontuais, são viáveis e benéficos para o consumidor do conjunto de mídias apresentado. / The multimedia consumption over the world wide web has been strongly growing along the last years. Hence it comes concerns about the bandwidth consumption, bolstered by significant raise of mobile computing. It is a known fact that multimedia bandwidth exceeded P2P network in terms of bandwidth consumption, and tend to grow in the next years. In this scenario, HTTP multimedia streaming technology came as a part of solution to the increasing multimedia bandwidth demand. MPEG-DASH is one of those technologies, currently employed at big distributing multimedia corporations, like Google and Netflix. However, the media transcoding and segmentation to MPEG-DASH standard is a laborious process and still slightly documented. This work cover in details techniques to media transcoding, segmentation, and reproduction on MPEG-DASH standard. The converted media are part of Centro de Educação a Distância of Universidade Federal de Juiz de Fora produced courseware. The work aims at technical process of MPEG DASH media standardization, with focus on raised problems, and includes some aspects of media reproduction. This work brings a case study, and an analysis of applicable aspects of computing environment. Were tested several techniques and tools and the problems, solutions and suggestions are reported through the work. Results shows that the results, although few problems, are viable and advantageous to consumer of the specified media set.

[pt] Este trabalho apresenta uma análise comparativa de algoritmos rápidos de estimação de movimento para codificações de vídeo, os quais visam reduzir a complexidade computacional do algoritmo Força Bruta. Os dois primeiros ( LOGD e 3 PASSOS) reduzem extremamente a complexidade, contudo, apresentam os mais baixos desempenhos. Dois deles Eliminação Sucessiva (ES I) e Adaptativo da Força Bruta (AFB) são técnicas recentes apresentadas na literatura. O dois últimos algoritmos (ES II e ES III) são modificações propostas nesta tese, com base nas técnicas ES I e AFB. Todos os algoritmos implementados neste trabalho são analisados isoladamente e em ambiente MPEG. / [en] This work presents a comparative analysis of fast motion compensation algorithms for vídeo compression, whitch aim at reducing the computacional complexity of the Full Search block matching tchnique. The first two ( LOG 2D and 3 Step) extremely reduce the complexity. However, they present the lowest performace. Two of them - Sucessive Elimination I ( ES I) and Adaptative Block Matching (AFB) - are schemes recently proposed in the literature. The last two algorithms (ES II and ES III) are modifications proposed in this thesis and are based on the ES I and AFB techniques. The algorithms are examined isolatedly an when operating in the MPEG environment.