We present the argument that video footage of real scenes can be used as input examples from which novel three-dimensional scenes can be created. We argue that the parameters used by traditional animation techniques based on the underlying physical properties of the water, do not intuitively relate to the resulting visual appearance. We will present a novel approach which allows a range of video examples to be used as a set of visual parameters to design the visible behaviour of a water animation directly. Our work begins with a method for reconstructing the perceived water surface geometry from video footage of natural scenes, captured with only a single static camera. We show that this has not been accomplished before, because previous approaches use sophisticated capturing systems which are limited to a laboratory environment. We will also present an approach for reconstructing the water surface velocities which are consistent with the reconstructed geometry. We then present a method of using these water surface reconstructions as building blocks which can be seamlessly combined to create novel water surface animations. We are also able to extract foam textures from the videos, which can be applied to the water surfaces to enhance their visual appearance. The surfaces we produce can be shaped and curved to fit within a user's three-dimensional scene, and the movement of external objects can be driven by the velocity fields. We present a range of results which show that our method can plausibly emulate a wide range of real-world scenes, different from those from which the water characteristics were captured. As the animations we create are fully three-dimensional, they can be rendered from any viewpoint, in any rendering style.

No description available.

Post-production facilities deal with large amounts of digital video, which presents difficulties when tracking, managing and searching this material. Recent research work in image and video analysis promises to offer help in these tasks, but there is a gap between what these systems can provide and what users actually need. In particular the popular research models for indexing and retrieving visual data do not fit well with how users actually work. In this thesis we explore how image and video analysis can be applied to an online video collection to assist users in reviewing and searching for material faster, rather than purporting to do it for them. We introduce a framework for automatically generating static 2-dimen- sional storyboards from video sequences. The storyboard consists of a series of frames, one for each shot in the sequence, showing the principal objects and motions of the shot. The storyboards are rendered as vector images in a familiar comic book style, allowing them to be quickly viewed and understood. The process consists of three distinct steps: shot-change detection, object segmentation, and presentation. The nature of the video material encountered in a post-production fa- cility is quite different from other material such as television programmes. Video sequences such as commercials and music videos are highly dy- namic with very short shots, rapid transitions and ambiguous edits. Video is often heavily manipulated, causing difficulties for many video processing techniques. We study the performance of a variety of published shot-change de- tection algorithms on the type of highly dynamic video typically encoun- tered in post-production work. Finding their performance disappointing, we develop a novel algorithm for detecting cuts and fades that operates directly on Motion-JPEG compressed video, exploiting the DCT coeffi- cients to save computation. The algorithm shows superior performance on highly dynamic material while performing comparably to previous algorithms on other material.

Η παρούσα ειδική ερευνητική εργασία εκπονήθηκε στα πλαίσια του Διατμηματικού Προγράμματος Μεταπτυχιακών Σπουδών Ειδίκευσης στα “Συστήματα Επεξεργασίας Σημάτων και Επικοινωνιών” στο Τμήμα Φυσικής του Πανεπιστημίου Πατρών. Αντικείμενο της παρούσας εργασίας είναι η σχεδίαση και ανάπτυξη του MPEG – 4 Simple Profile CODEC στο περιβάλλον Simulink με σκοπό την τελική εκτέλεση του αλγορίθμου DSP που θα προκύψει, στην πλατφόρμα ανάπτυξης TMS320DM6437 EVM. Στο πρώτο κεφάλαιο ορίζεται η έννοια της κωδικοποίησης βίντεο σε πραγματικό χρόνο και περιγράφεται η σύγχυση που επικρατεί γύρω από αυτήν. Επίσης γίνεται μια περιγραφή των επεξεργαστών ψηφιακού σήματος ως προς τα τυπικά χαρακτηριστικά που διαθέτουν, την αρχιτεκτονική τους, την αρχιτεκτονική μνήμης, τα στοιχεία υλικού που διαθέτουν για τη ροή του DSP προγράμματος, ενώ παράλληλα, παρουσιάζεται η ιστορική εξέλιξη των DSPs που οδήγησε στους σύγχρονους DSPs και οι οποίοι, διαθέτουν καλύτερες επιδόσεις από τους προπάτορές τους, και αυτό χάρη στις τεχνολογικές και αρχιτεκτονικές εξελίξεις όπως, οι χαμηλότεροι κανόνες σχεδίασης, η γρήγορη προσπέλαση κρυφής μνήμης δύο επιπέδων, η σχεδίαση του DMA και ενός μεγαλύτερου συστήματος διαύλου. Στο τέλος του κεφαλαίου παρουσιάζεται η αρχιτεκτονική της πλατφόρμας ανάπτυξης TMS320DM6437 EVM καθώς και οι διεπαφές υλικού που διαθέτει για την είσοδο και έξοδο βίντεο/ήχου από αυτήν. Στο δεύτερο κεφάλαιο γίνεται μια εκτενής παρουσίαση των εννοιών που συναντώνται στην κωδικοποίηση βίντεο. Στην αρχή του κεφαλαίου απεικονίζεται το γενικό μοντέλο ενός κωδικοποιητή/αποκωδικοποιητή και βάσει αυτού προχωράμε στην περιγραφή του χρονικού μοντέλου, το οποίο επιβάλλει την πρόβλεψη του τρέχοντος πλαισίου βίντεο χρησιμοποιώντας το προηγούμενο, ενώ παράλληλα, εξηγεί και μεθόδους για την εκτίμηση κίνησης περιοχών (μακρομπλοκ) μέσα στο πλαίσιο ενός βίντεο και το πώς μπορεί να γίνει ο υπολογισμός του σφάλματος κίνησης τους. Στη συνέχεια περιγράφεται το μοντέλο εικόνας το οποίο στην πράξη αποτελείται από τρία συστατικά μέρη: τον μετασχηματισμό (αποσυσχετίζει και συμπιέζει τα δεδομένα), την κβάντιση (μειώνει την ακρίβεια των μετασχηματισμένων δεδομένων) και την ανακατάταξη (ανακατατάσσει τα δεδομένα ούτως ώστε να ομαδοποιήσει μαζί τις σημαντικές τιμές). Οι συντελεστές του μετασχηματισμού μετά την ανακατάταξη και την κωδικοποίηση, μπορούν να κωδικοποιηθούν περαιτέρω με τη χρήση κωδικών μεταβλητού μήκους (Huffman κωδικοποίηση) ή μέσω αριθμητικής κωδικοποίησης. Στο τέλος του κεφαλαίου περιγράφεται το υβριδικό μοντέλο DPCM/DCT CODEC πάνω στον οποίο στηρίζεται και η υλοποίηση του MPEG – 4 Simple Profile CODEC. Στο τρίτο κεφάλαιο ουσιαστικά γίνεται μια περιγραφή των χαρακτηριστικών του MPEG – 4 Simple Profile CODEC, των εργαλείων που χρησιμοποιεί, της έννοιας αντικείμενο που πλέον υπεισέρχεται στην κωδικοποίηση βίντεο καθώς και τα είδη προφίλ και επιπέδων που υποστηρίζει το συγκεκριμένο πρωτόκολλο κωδικοποίησης/αποκωδικοποίησης. Στο τέταρτο κεφάλαιο παρουσιάζεται η υλοποίηση του κωδικοποιητή, του αποκωδικοποιητή του MPEG – 4 Simple Profile CODEC καθώς και των επιμέρους υποσυστημάτων που τους απαρτίζουν. Στο πέμπτο κεφάλαιο περιγράφεται η αλληλεπίδραση του χρήστη με το σύστημα κωδικοποίησης/αποκωδικοποίησης, τι παράμετροι χρειάζονται να δοθούν ως είσοδοι από αυτόν, καθώς και πως είναι δυνατή η χρήση του συγκεκριμένου συστήματος. / This project objective is the design and development of MPEG – 4 Simple Profile CODEC in Simulink environment in order to execute the resulting DSP algorithm on the development platform TMS320DM6437 EVM. The first chapter defines the term of real – time video coding which sometimes is misunderstood by most people. Besides there is a brief description of DSP systems, which includes information about their typical characteristics, their architecture, their memory architecture and the hardware elements provided with in order to support the flow of a DSP program. It is also presented the evolution of DSPs through time, which finally gave the modern DSPs with better performance than their ancestors thanks to the technological and architectonical improvements such as, lower design rules, fast-access two-level cache, (E)DMA circuitry and a wider bus system. At the end of this chapter it is presented the architecture of TMS320DM6437 EVM board and its input/output hardware interfaces for video and sound. At the second chapter there is an extensive presentation of terms found at the science of coding/decoding video. At the beginning of this chapter it is depicted a general model including a video encoder/decoder and this is the reason for the description of temporal model, which includes the prediction of current frame from the previous one, and at the same time it explains the computation methods of macroblock motion estimation and motion compensation. Continuing it is described the image model aparted from three component parts, the transformation (decorrelation and data compression), the quantization (reduces the accuracy of transformed data) and the reordering (reorders data on a way that groups significant values all together). The transform coefficients after reordering and coding, can be further coding by using variable length coding (Huffman coding) or arithmetic coding. At the end of the chapter the hybrid model of DPCM/DCT CODEC is described and this is the one where the implementation of MPEG – 4 Simple Profile CODEC has been set up. At the third chapter there is a description about the characteristics of MPEG – 4 Simple Profile CODEC, the tools used, the “object” term, which appears on video coding/decoding and also what are the profiles and levels supported by the specific video encoding/decoding protocol. Finally it is described how the coding of rectangular frames is done and the Simulink model of MPEG – 4 Simple Profile CODEC which is the base for the implementation of DSP algorithm executed on the development platform. At the forth chapter we present the implementation of MPEG – 4 Simple Profile CODEC encoder/decoder and their partial subsystems. At the fifth chapter it is described the interaction between user and the CODEC, what are the parameters needed to be entered as inputs and how the system can be used.

In digital video compression, apart from storage, successful transmission of the compressed video data over the bandwidth limited erroneous channels is another important issue. To enable a video codec for broadcasting application, it is required to implement the corresponding coding tools (e.g. error-resilient coding, rate control etc.). They are normally non-normative parts of a video codec and hence their specifications are not defined in the standard. In Dirac as well, the original codec is optimized for storage purpose only and so, several non-normative part of the encoding tools are still required in order to be able to use in other types of application. Being the "Research and Developments of the Dirac Video Codec" as the research title, phase I of the project is mainly focused on the error-resilient transmission over a noisy channel. The error-resilient coding method used here is a simple and low complex coding scheme which provides the error-resilient transmission of the compressed video bitstream of Dirac video encoder over the packet erasure wired network. The scheme combines source and channel coding approach where error-resilient source coding is achieved by data partitioning in the wavelet transformed domain and channel coding is achieved through the application of either Rate-Compatible Punctured Convolutional (RCPC) Code or Turbo Code (TC) using un-equal error protection between header plus MV and data. The scheme is designed mainly for the packet-erasure channel, i.e. targeted for the Internet broadcasting application. But, for a bandwidth limited channel, it is still required to limit the amount of bits generated from the encoder depending on the available bandwidth in addition to the error-resilient coding. So, in the 2nd phase of the project, a rate control algorithm is presented. The algorithm is based upon the Quality Factor (QF) optimization method where QF of the encoded video is adaptively changing in order to achieve average bitrate which is constant over each Group of Picture (GOP). A relation between the bitrate, R and the QF, which is called Rate-QF (R-QF) model is derived in order to estimate the optimum QF of the current encoding frame for a given target bitrate, R. In some applications like video conferencing, real-time encoding and decoding with minimum delay is crucial, but, the ability to do real-time encoding/decoding is largely determined by the complexity of the encoder/decoder. As we all know that motion estimation process inside the encoder is the most time consuming stage. So, reducing the complexity of the motion estimation stage will certainly give one step closer to the real-time application. So, as a partial contribution toward realtime application, in the final phase of the research, a fast Motion Estimation (ME) strategy is designed and implemented. It is the combination of modified adaptive search plus semi-hierarchical way of motion estimation. The same strategy was implemented in both Dirac and H.264 in order to investigate its performance on different codecs. Together with this fast ME strategy, a method which is called partial cost function calculation in order to further reduce down the computational load of the cost function calculation was presented. The calculation is based upon the pre-defined set of patterns which were chosen in such a way that they have as much maximum coverage as possible over the whole block. In summary, this research work has contributed to the error-resilient transmission of compressed bitstreams of Dirac video encoder over a bandwidth limited error prone channel. In addition to this, the final phase of the research has partially contributed toward the real-time application of the Dirac video codec by implementing a fast motion estimation strategy together with partial cost function calculation idea.

Αντικείμενο της παρούσας διπλωματικής εργασίας είναι αρχικά η καταγραφή της ανθρώπινης κίνησης με κάποια συσκευή παρακολούθησης και κατόπιν η δημιουργία ενός αντιπροσωπευτικού μοντέλου, ώστε να μπορεί να μελετηθεί η δυναμική του συμπεριφορά. Ως συσκευή καταγραφής χρησιμοποιήθηκε ο αισθητήρας Kinect της Microsoft. Το μοντέλο που αναπτύχθηκε αφορά κυρίως τα κάτω άκρα του ανθρώπου και επιπλέον διαθέτει μυοσκελετική δομή με 86 μύες. Στα πλαίσια των αναλύσεων χρησιμοποιήθηκαν διάφορες τεχνικές για την εξαγωγή των αποτελεσμάτων, όπως είναι η αντίστροφη κινηματική, αντίστροφη δυναμική, υπολογισμός μυϊκών διεγέρσεων και ορθή δυναμική και προτείνουμε μια στρατηγική για την ανάλυση και την εξαγωγή αποτελεσμάτων. / The research developed in this thesis first deal with the problem of capturing the human body motion and then concentrates on the creation of musculoskeletal models, which can capture and accurately study its dynamical behavior. The Microsoft's Kinect sensor was utilized to capture the human motion. The model used for the simulations is the human lower extremity with 86 attached muscles. For the analysis phase we used some common methods such as inverse kinematics, inverse dynamics, computed muscle control and forward dynamics and we showed a general pipeline strategy for generating correct results.

Αντικείμενο της παρούσας διπλωματικής εργασίας είναι η παρουσίαση και η μελέτη ενος εναλλακτικού σχεδιασμού του deblocking φίλτρου του προτύπου κωδικοποίησης βίντεο Η.264. Αρχικά επεξηγείται αναλυτικά ο τρόπος λειτουργίας του φίλτρου και στη συνέχεια προτείνεται ένας πρωτοποριακός σχεδιασμός με χρήση pipeline πέντε σταδίων. Ο σχεδιασμός παρουσιάζει σημαντικά πλεονεκτήματα στον τομέα της ταχύτητας (ενδεικτικά εμφανίζεται βελτιωμένη απόδοση στην συχνότητα λειτουργίας και στο throughput). Αυτό πιστοποιήθηκε από μετρήσεις που έγιναν σε συγκεκριμένα fpga και επαλήθευσαν τα θεωρητικά συμπεράσματα που είχαν εξαχθεί. / The standard H.264 (or else MPEG-4 part 10) is nowadays the most widely used standard in the area of video coding as it is supported by the largest enterprises in the internet (including Google, Apple and Youtube). Its most important advantage over the previous standards is that it achieves better bitrate without falling in terms of quality. A crucial part of the standard is the deblocking filter which is applied in each macroblock of a frame so that it reduces the blocking distortion. The filter accounts for about one third of the computational requirements of the standard, something which makes it a really important part of the filtering process. The current diploma thesis presents an alternative design of the filter which achieves better performance than the existing ones. The design is based in the use of two filters (instead of one used in current technology) and moreover, in the application of a pipelined design in each filter. By using a double filter, exploitation of the independence which exists in many parts of the macroblock is achieved. That is to say, it is feasible that different parts of it can be filtered at the same time without facing any problems. Furthermore, the use of the pipeline technique importantly increases the throughput. Needless to say, in order for the desired result to be achieved, the design has to be made really carefully so that the restrictions imposed by the standard will not be failed. The use of this alternative filter design will result in an important raise in the performance. Amongst all, the operating frequency, the throughput and the quality of the produced video will all appear to be considerably risen. It also needs to be mentioned that the inevitable increase of the area used (because of the fact that two filters are used instead of one) is not really important in terms of cost. The structure of the thesis is described in this paragraph. In chapter 1 there is a rather synoptic description of the H.264 standard and the exact position of the deblocking filter in the whole design is clarified. After that, the algorithmic description of the filter follows (Chapter 2). In this chapter, all the parameters participating in the filter are presented in full detail as well as the equations used during the process. In the next chapter (chapter 3), the architecture chosen for the design is presented. That is to say, the block diagram is presented and explained, as well as the table of timings which explains completely how the filter works. The pipelining technique applied in the filter is also analyzed and justified in this chapter. In the next chapter (chapter 4), every structural unit used in the current architecture is analyzed completely and its role in the whole structure is presented. Finally, in chapter 5, the results of the measurements made in typical fpgas of Altera and Xilinx are presented. The results are shown in table format whereas for specific parameters diagrams were used so that the improved performance of the current design compared to the older ones that are widely used, becomes evident.

Στην παρούσα εργασία υλοποιούμε ένα σύστημα δημιουργίας περιλήψεων από ακολουθίες βίντεο. Υλοποιούνται όλα τα βήματα που θα πρέπει να ακολουθηθούν (εξαγωγή χαρακτηριστικών-ανίχνευση πλάνων-εξαγωγή χαρακτηριστικών καρέ) έτσι ώστε να εξαχθεί ένα σύνολο καρέ (χαρακτηριστικά καρέ) τα οποία να συνοψίζουν νοηματικά το περιεχόμενο μιας ακολουθίας βίντεο. Η επεξεργασία του βίντεο γίνεται απευθείας στο συμπιεσμένο πεδίο και συγκεκριμένα σε συμπιεσμένα αρχεία MPEG-1-2, έτσι ώστε τα αποτελέσματα να εξάγονται σε σχετικά μικρό χρόνο και με σχετικά χαμηλές απαιτήσεις σε αποθηκευτικό χώρο και επεξεργαστική ισχύ. / In this paper a video summarization system is being constructed. We acomplish all the needed steps (feature extraction -shot detection-keyframe extraction) in order to extract a set of frames (keyframes) that capture the semantic content of the video sequence. The processing of the video takes place directly at the compressed domain (at MPEG-1-2 video files). Thus we obtain results at relatively little time and with relatively low storage and computer power demands.

L'imagerie numérique, de la synthèse d'images à la vision par ordinateur est en train de connaître une forte évolution, due entre autres facteurs à la démocratisation et au succès commercial des caméras 3D. Dans le même contexte, l'impression 3D grand public, qui est en train de vivre un essor fulgurant, contribue à la forte demande sur ce type de caméra pour les besoins de la numérisation 3D. L'objectif de cette thèse est d'acquérir et de maîtriser un savoir-faire dans le domaine de la capture/acquisition de modèles 3D en particulier sur l'aspect rendu réaliste. La réalisation d'un scanner 3D à partir d'une caméra RGB-D fait partie de l'objectif. Lors de la phase d'acquisition, en particulier pour un dispositif portable, on est confronté à deux problèmes principaux, le problème lié au référentiel de chaque capture et le rendu final de l'objet reconstruit. / Digital imaging, from the synthesis of images to computer vision isexperiencing a strong evolution, due among other factors to the democratization and commercial success of 3D cameras. In the same context, the consumer 3D printing, which is experiencing a rapid rise, contributes to the strong demand for this type of camera for the needs of 3D scanning. The objective of this thesis is to acquire and master a know-how in the field of the capture / acquisition of 3D models in particular on the rendered aspect. The realization of a 3D scanner from a RGB-D camera is part of the goal. During the acquisition phase, especially for a portable device, there are two main problems, the problem related to the repository of each capture and the final rendering of the reconstructed object.