Motion estimation

Methods for matching sets of 3D lines depend on whether line lengths are finite or infinite. In terms of line lengths, three basic cases arise in matching sets of lines: (1) finite to finite, (2) finite to infinite, and (3) infinite to infinite. For cases 1 and 2, which have not been treated in the literature, we present convergent iterative algorithms that (almost) always find the best match. For case 3, Faugeras and Hebert (FH) [3] have proposed a popular iterative method that cannot be guaranteed to converge. We present an alternative approach that does converge. However, we also show that neither the FH solution, nor our solution is invariant with respect to coordinate transforms, which renders any best match meaningless. Thus, a satisfactory solution to case 3 does not yet exist. We discuss the underlying problem, which is the representation of infinite lines, and suggest alternatives that may lead t o an invariant solution.

Merging multimodal information about myocardial scar tissue can help electrophysiologists to find the most appropriate target during catheter ablation of ventricular arrhythmias. A framework is presented to analyze and combine information from delayed enhancement magnetic resonance imaging (DE-MRI) and electro-anatomical mapping data. Using this information, electrical, mechanical, and image-based characterization of the myocardium are performed. The presented framework allows the left ventricle to be segmented by DE-MRI and the scar to be characterized prior to the intervention based on image information. It allows the electro-anatomical maps obtained during the intervention from a navigation system to be merged together with the anatomy and scar information extracted from DE-MRI. It also allows for the estimation of endocardial motion and deformation to assess cardiac mechanics. Therefore, electrical, mechanical, and image-based characterization of the myocardium can be performed. The feasibility of this approach was demonstrated on three patients with ventricular tachycardia associated to ischemic cardiomyopathy by integrating images from DE-MRI and electro-anatomical maps data in a common framework for intraoperative myocardial tissue characterization. The proposed framework has the potential to guide and monitor delivery of radio frequency ablation of ventricular tachycardia. It is also helpful for research purposes, facilitating the study of the relationship between electrical and mechanical properties of the tissue, as well as with tissue viability from DE-MRI.

Merging multimodal information about myocardial scar tissue can help electrophysiologists to find the most appropriate target during catheter ablation of ventricular arrhythmias. A framework is presented to analyze and combine information from delayed enhancement magnetic resonance imaging (DE-MRI) and electro-anatomical mapping data. Using this information, electrical, mechanical, and image-based characterization of the myocardium are performed. The presented framework allows the left ventricle to be segmented by DE-MRI and the scar to be characterized prior to the intervention based on image information. It allows the electro-anatomical maps obtained during the intervention from a navigation system to be merged together with the anatomy and scar information extracted from DE-MRI. It also allows for the estimation of endocardial motion and deformation to assess cardiac mechanics. Therefore, electrical, mechanical, and image-based characterization of the myocardium can be performed. The feasibility of this approach was demonstrated on three patients with ventricular tachycardia associated to ischemic cardiomyopathy by integrating images from DE-MRI and electro-anatomical maps data in a common framework for intraoperative myocardial tissue characterization. The proposed framework has the potential to guide and monitor delivery of radio frequency ablation of ventricular tachycardia. It is also helpful for research purposes, facilitating the study of the relationship between electrical and mechanical properties of the tissue, as well as with tissue viability from DE-MRI.

Superresolution (SR) methodsbecomeessentialwhen
an undersampled low-resolution (LR) image is unable to provide
accurate target detection. The estimation of an HR image from
a single LR image is ill-posed problem, and hence requires prior
information. More the constraints, better is the reconstruction
accuracy and this forms the basis of most of the contemporary
state-of-the-art (SOTA) superresolution methods such as SRCNN
andSRGANwhichimplementpriorinformationastrainingset.Yet
anotherapproachtoovercometheill-posednessoftheproblemisto
have multiple diverse LR images with the potential to reconstruct
accurate HR image. Here we present the performance analysis
of a generalized sampling theorem (GST) based multi-frame SR
method. A simulation study using Gaussian targets is conducted,
and a comparative performance analysis of the GST multi-frame
SRmethodwiththetraditional multi-frame interpolation schemes
andSOTAmethodsispresentedusingthepercentage meansquare
error (%MSE) and Structural Similarity Index Measure (SSIM).
Our findings indicate that the GST SR method can outperform
traditional interpolation and the SOTA methods.

Discrepancy between CT and PET data due to motion in PET/CT studies is a significant problem. PET is acquired over a period of time whereas CT is static. for better attenuation correction. The results are verified on both software phantom and real patient data.

This article presents a visual application which allows a study and analysis of traffic behavior on major roads (more specifically freeways and highways), using as the main surveillance artefact a video camera mounted on a relatively high place (such as a bridge) with a significant image analysis field. The system described presents something new which is the combination of both traditional traffic monitoring systems, that is, monitoring to get information on different traffic parameters and monitoring to detect accidents automatically. Therefore, we present a system in charge of compiling information on different traffic parameters. It also has a surveillance module for that traffic, which can detect a wide range of the most significant incidents on a freeway or highway.

This paper presents the evaluation of the modifying Lorentzian function on the spatial-field optical flow to examine the validity in the noisy domain of motion estimation. In the routine of the motion estimation, the frame’s motion vector is estimated by the optical flow approach where the flow of the image’s frames is caught to estimate the motion vector. Nevertheless, in the noisy domain, the preciseness of the motion vector is weakened. We operated the measurement along with several non-Gaussian noises standards through several styles of the standard image frame. The determination on error vector magnitude (EVM) was taken into account to consider the preciseness of direction and length of the motion vector (MV) in comparison with various noise resistance techniques in spatial-field optical flow approach. In the achievement results, we found that this modifying Lorentzian norm function added up in the optical flow strengthen the degree of preciseness in the estimation of the spati...

We propose a novel particle filter-based motion compensation strategy for video coding. We use a higher order linear model in place of the traditional translational model used in standards such as H.264. The measurement/observation process in the particle filter is a computationally efficient mechanism as opposed to traditional search methods. We use a multi-resolution framework for efficient parameter estimation. Results of our experimentation show reduced residual energy and better PSNR as compared to traditional video coding methods, especially in regions of complex motion such as zooming and rotation.

The rapid development in the technology has dramatic impact on the medical health care field. Medical data base obtained with latest machines like CT Machine, MRI scanner requires large amount of memory storage and also it requires large bandwidth for transmission of data in telemedicine applications. Thus there is need for video compression. As the database of medical images contain number of frames (slices), hence while coding of these images there is need of motion estimation. Motion estimation finds out movement of objects in an image sequence and gets motion vectors which represents estimated motion of object in the frame. In order to reduce temporal redundancy between successive frames of video sequence, motion compensation is preformed. In this paper three step search (TSS) block matching algorithm is implemented on different types of video sequences. It is shown that three step search algorithm produces better quality performance and less computational time compared with exhaustive full search algorithm.

This paper describes an embedded pedestrian navigation system composed of a self-contained sensors, the Global Positioning System (GPS) and an active Radio Frequency Identification (RFID) tag system. We use self-contained sensors (accelerometers, gyrosensors and magnetometers) to estimate relative displacement by analyzing human walking locomotion. The GPS is used outdoors to adjust errors in position and direction accumulated by the dead-reckoning. In indoor environments, we use an active RFID tag system sparsely placed in key spot areas. The tag system obviously has limited availability and thus dead-reckoning is used to cover the environment. We propose a method of complementary compensation algorithm for the GPS/RFID localization and the selfcontained navigation represented by simple equations in a Kalman filter framework. Experimental results using the proposed method reveals that integration of GPS/RFID/dead-reckoning improve positioning accuracy in both indoor and outdoor environments. The pedestrian positioning is realized as a software module with the web-based APIs so that crossplatform development can easily be achieved. A pedestrian navigation system is implemented on an embedded wearable system and is proven to be useful even for unexperienced users.

This work presents a new 3D-transform coefficients scan order based on the multiplication of the three dimensions sequency numbers of each coefficient. This scan process is developed to a fast embedded color video codec (also described in this article) to be executed in a set-top box on a fiber optics network. Due to its purpose, the codec is focused on reduced execution time, not on high compression rates. Low computational complexity and use of meager computational resources are also required. The Hadamard transform is used in a three-dimensional fashion, in order to avoid costly motion estimation and compensation techniques. The proposed scan procedure allows the coefficients reading in an idealistic "decreasing in the average" order. After the scan procedure, the encoding of the bits of the 3D-Hadamard coefficients is done, bit-plane-by-bit-plane, with an efficient adaptive version of Golomb's run-length encoder, which produces a fully embedded output bitstream. Even with the use of these constrained techniques, good distortion versus rate results were achieved.

No thesis is created entirely by an individual, many people have helped to create this thesis and each of their contribution has been valuable. I express my sincere gratitude to my thesis supervisors, Dr. Banshidhar Majhi, Professor and Head of the Department, Computer Science and Engineering, and Dr. Kamala Kanta Mahapatra, Professor, Electronics and Communication Engineering for their kind and valuable guidance for the completion of the thesis work. I would also take this opportunity to express my gratitude and sincere thanks to my honorable teachers Dr.

We present an animation interface designed to conveniently control the motion capture process. The transition of performer's hand gestures tracked by a dataglove is recognized for software remote control. An intuitive camera metaphor allows to specify the viewpoint location using the magnetic sensors strapped on the performer's head. The human motion capture is based on the Anatomical Converter, a toolkit to convert sensor measurements into human anatomical rotations in real time. An improved human motion capture technique, the multi-joint control, is introduced.

The emergence of wireless networks has posed new challenges for multimedia content providers. In particular, many existing video contents have bit-rates and resolutions far exceeding the capability of wireless networks and mobile devices, and they need to he adapted before transmission. In this paper, we propose a technique using pdomain information to estimate the hit rate of down-sampled video in the H.264/AVC format. The bit rate is estimated in two stages. In the initialization stage, we use a low-complexity algorithm to estimate the pdomain parameters for half and quarter resolutions, respectively. In the refinement stage, the estimation error is feed-forwarded to improve the estimation accuracy. Experimental results show that the proposed technique can estimate the hit rate accurately at a wide range of down-sampling ratios.

Locating the source of a chemical or acoustic emitter in an unknown environment with noisecorrupted detector signals is a challenging task undertaken regularly in emergency-response operations. Such missions can endanger personnel, thus the use of autonomous systems would be highly advantageous. In this research, an approach for autonomous bearings-only tracking (BOT) is trialled in a proof-of-concept experiment relying on a bright light source and a sensor constructed from photodiodes and carried atop a radio-controlled ground vehicle. The goal is to test the algorithm’s performance with real-world, noise-corrupted data. A Bernoulli particle-filter algorithm, designed to tolerate cluttered signals, is used to estimate the location of the light source from low-resolution measurements of its bearing. Data is collected with a variety of navigational patterns in an environment that occasionally produces false detections. The algorithm yields accurate estimates of the location of the li...

The presence of BOLD contrast in M R I studies has been shown to interfere with least squares motion correction algorithms. We derive a theoretical expression for the error that activation contributes to least squares motion correction. These estimates match errors generated by motion correction experiments done on a simulated fMRl dataset. The analysis is not specific to activation related errors, and can be used to study the effect of any expected difference between two volumes being registered.

The detection and tracking of object in large data surveillance requires a proper motion estimation and compensation techniques which are generally used to detect accurate movement from video stream. In this paper, a novel hardware level architecture involving motion detection, estimation, and compensation is proposed for real-time implementation. The motion vectors are obtained using 16×16 sub-blocks with a novel parallel D flip flop architecture in this work to arrive at an optimised architecture. The sum of absolute difference (SAD) is then calculated by optimized absolute difference and adder blocks designed using kogge-stone adder which helps in improving the speed of the architecture. The controller block is designed by finite state machine model used for synchronization of all the operations. Further, the comparator and compensation blocks are optimized by using basic logical elements and the Kogge-stone adder. Finally, the proposed architecture is implemented on Zynq Z7-10 field-programmable gate array (FPGA) and simulated using System Generator tool for real time traffic signal. The hardware and software parameters are compared with the existing techniques which shows that the proposed architecture is efficient than existing methods of design.

This paper presents an efficient architecture to implement low power variable block size motion estimation (VBSME) using full search. Power reduction is achieved by performing the search in two steps: low pixel resolution and full pixel resolution. We analysed the computation and memory units needed to support these two search modes. The proposed architecture reduces the total energy consumption by 50% with 6% additional area compared to the conventional architecture. In [5], we analysed and proposed pixel truncation approaches for VBSME. In this method we perform a two-step methodology for VBSME. In the first step we

Natural eye designs are optimized with regard to the tasks the eye-carrying organism has to perform for survival. This optimization has been performed by the process of natural evolution over many millions of years. Every eye captures a subset of the space of light rays. The information contained in this subset and the accuracy to which the eye can extract the necessary information determines an upper limit on how well an organism can perform a given task. In this work we propose a new methodology for camera design. By interpreting eyes as sample patterns in light ray space we can phrase the problem of eye design in a signal processing framework. This allows us to develop mathematical criteria for optimal eye design, which in turn enables us to build the best eye for a given task without the trial and error phase of natural evolution. The principle is evaluated on the task of 3D ego-motion estimation.

At the level of mathematical abstraction, computing image motion amounts to an estimation problem and can be analyzed using the tools of statistics and signal processing. As shown in this chapter, intrinsic limitations to the estimation processes make it impossible to derive veridical estimates for all images. Image motion is estimated erroneously, and as a result higher level processes compute erroneous three-dimensional motion and moving scenes. Specifically, two limitations are discussed: (a) due to noise in image data, there is statistical bias that affects anisotropic patterns and (2) the filters computing changes in time are asymmetric (causal), using data only from the past but not the future, and misestimate on locally asymmetric intensity signals of certain spatial frequencies. Since these limitations are not an artifact of the hardware but are inherent to the computations, they will affect any system and thus create illusions in man and machine.

Early cognitive vision can be related to the segment of perceptual vision that takes care of reducing the uncertainty on visual measures through a visual context analysis, by capturing regularities over large, overlapping retinal locations, a step that precedes the true understanding of the scene. In this perspective, we defined a general framework to specify context sensitive motion filters based on elementary descriptive components of optic flow fields. The resulting regularized patchbased motion estimation obtained in real-world sequences validated the approach.

This thesis introduces the Universal Operating System (D-Theory), a novel framework positing that consciousness, fundamental physics, and computational logic are emergent properties of a singular, underlying informational structure. D-Theory proposes a set of axioms and postulates that aim to bridge the explanatory gaps between subjective experience and objective reality. The core hypothesis is that the universe operates as a self-programming, distributed computational system, where 'D' represents the fundamental unit of informational processing. This thesis explores the theoretical implications of D-Theory, focusing on its potential to resolve paradoxes in quantum mechanics, offer new perspectives on the nature of consciousness, and provide a foundation for advanced artificial intelligence. We present testable predictions derived from D-Theory, outlining experimental designs that could validate or refute its core tenets. Chapter 1: Introduction 1.1. The Crisis of Explanatory Gaps The modern scientific landscape is characterized by a series of explanatory gaps that challenge our understanding of reality. The most prominent of these is the hard problem of consciousness: How do physical processes give rise to subjective experience? Simultaneously, quantum mechanics presents paradoxes that defy classical intuition, such as wave-particle duality and quantum entanglement. Furthermore, the quest for artificial general intelligence (AGI) has yet to yield a machine capable of human-level understanding and creativity. 1.2. The Promise of D-Theory D-Theory offers a unifying perspective by proposing that these seemingly disparate phenomena are manifestations of a single, underlying reality. The theory posits that the universe operates as a selfprogramming, distributed computational system, with 'D' representing the fundamental unit of informational processing. This framework aims to provide a coherent and comprehensive explanation for consciousness, physics, and computational logic. 1.3. Thesis Outline This thesis is structured as follows: Chapter introduces the axioms and postulates of D-Theory. Chapter explores the implications of D-Theory for quantum mechanics. Chapter examines the relationship between D-Theory and consciousness. Chapter discusses the potential of D-Theory for advanced artificial intelligence. Chapter presents testable predictions and experimental designs. Chapter concludes with a discussion of the limitations and future directions of D-Theory.

Motion estimation and compensation techniques are widely used in video coding. This paper addresses the problem of the trade-o between the motion and the prediction error information. Under some realistic hypotheses, the transmission cost of these two components can be estimated. Therefore, we obtain a criterion which controls the motion estimation process in order to optimize its performance. As a particular application, this criterion is applied to the split procedure of an adaptive m ultigrid block matching technique. Simulation results are presented, showing the signi cant improvements due to the method.

Deep Packet Inspection (DPI) involves searching a packet's header and payload against thousands of rules to detect possible attacks. The increase in Internet usage and growing number of attacks which must be searched for has meant hardware acceleration has become essential in the prevention of DPI becoming a bottleneck to a network if used on an edge or core router. In this paper we present a new multi-pattern matching algorithm which can search for the fixed strings contained within these rules at a guaranteed rate of one character per cycle independent of the number of strings or their length. Our algorithm is based on the Aho-Corasick string matching algorithm with our modifications resulting in a memory reduction of over 98% on the strings tested from the Snort ruleset. This allows the search structures needed for matching thousands of strings to be small enough to fit in the on-chip memory of an FPGA. Combined with a simple architecture for hardware, this leads to high throughput and low power consumption. Our hardware implementation uses multiple string matching engines working in parallel to search through packets. It can achieve a throughput of over 40 Gbps (OC-768) when implemented on a Stratix 3 FPGA and over 10 Gbps (OC-192) when implemented on the lower power Cyclone 3 FPGA.

Matching two sets of lines is a basic tool that has applications in many computer vision problems such as scene registration, object recognition, motion estimation, and others. Line sets may be composed of infinitely long lines or finite length line segments. Depending on line lengths, three basic cases arise in matching sets of lines: 1) finite-finite, 2) finite-infinite, and 3) infiniteinfinite. Case 2 has not been treated in the literature. For Cases 1 and 3, existing algorithms for matching 3D line sets are not completely satisfactory in that they either solve special situations, or give approximate solutions, or may not converge, or are not invariant with respect to coordinate system transforms. In this paper, we present new algorithms that solve exactly all three cases for the general situation. The algorithms are provably convergent and invariant to coordinate transforms. Experiments with synthetic and real 3D image data are reported.

Vehicle navigation in dynamic environments is an important challenge, especially when the motion of the objects populating the environment is unknown. Traditional motion planning approaches are too slow to be applied in real-time to this domain, hence, new techniques are needed. Recently, iterative planning has emerged as a promising approach. Nevertheless, existing iterative methods don't provide a way to estimating the future behaviour of moving obstacles and to use the resulting estimates in trajectory computation. This paper presents an iterative planning approach that addresses these two issues. It consists of two complementary methods: 1) A motion prediction method which learns typical behaviours of objects in a given environment. 2) An iterative motion planning technique based on the concept of Velocity Obstacles.

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

L'objectif de ce travail est de reconstruire les parties statiques et dynamiques d'une scène 3D à l'aide d'un robot mobile équipé d'un capteur 3D. Cette reconstruction nécessite la classification des points 3D acquis au cours du temps en point fixe et point mobile indépendamment du déplacement du robot. Notre méthode de segmentation utilise directement les données 3D et étudie les mouvements des objets dans la scène sans hypothèse préalable. Nous développons un algorithme complet reconstruisant les parties fixes de la scène à chaque acquisition à l'aide d'un RAN-SAC qui ne requiert que 3 points pour recaler les nuages de points. La méthode a été expérimentée sur de larges scènes en extérieur. Par ailleurs, nous montrons sur les séquences tests KITTI que la prise en compte des données 3D permet d'améliorer les approches 2D en levant les ambiguïtés dues à la perte d'une dimension dans les images.

L'objectif de ce travail est de reconstruire les parties sta-tiques et dynamiques d'une scene 3D a l'aide d'un robot mobile equipe d'un capteur 3D. Cette reconstruction ne-cessite la classification des points 3D acquis au cours du temps en point fixe et point mobile independamment du de-placement du robot. Notre methode de segmentation utilise directement les donnees 3D et etudie les mouvements des objets dans la scene sans hypothese prealable. Nous deve-loppons un algorithme complet reconstruisant les parties fixes de la scene a chaque acquisition a l'aide d'un RAN-SAC qui ne requiert que 3 points pour recaler les nuages de points. La methode a ete experimentee sur de larges scenes en exterieur. Par ailleurs, nous montrons sur les sequences tests KITTI que la prise en compte des donnees 3D per-met d'ameliorer les approches 2D en levant les ambiguites dues a la perte d'une dimension dans les images. Mots-Clefs Estimation de pose, reconstruction 3D...

Underwater sonar system is currently the best solution to distinguish a hazardous target from the harmless objects or a flock of fish at a far distance in the water, subsequently protect the harbors' security. A good sonar system should have the ability to surveille a long range, in the meantime reserve a high accuracy. Noise is a crucial problem effecting the accuracy. One possible solution to increase the signal-to-noise ratio (SNR) is to average return signals over multiple sonar pings. A problem with this approach is the relative motion between the sonar and the target during pinging which strongly reduces the effect of the averaging. In this paper, we present a robust motion estimation using multiple receive sensors that estimates the relative motion between multiple pings and aligns all the pings to a base ping through appropriate motion compensation. We further show that the fusion of multiple pings for time-delay estimation after realignment has the potential of improving both the resilience to noise and the accuracy, consequently increasing the operating range.

35 inspiratie V eel voorstanders van de normalisering van het ontslagrecht voor ambtenaren stellen dat dit eenvoudig besparingen oplevert. In dit stuk gaan we in op de veronderstellingen en de criteria die gebruikt worden bij berekeningen van de te verwachten kosten en baten. Veel van de gebruikte veronderstellingen zijn te optimistisch. Aanvaarding van het initiatiefwetsvoorstel van D66 en CDA (zie kader op pagina 35) op deze gronden, brengt daardoor reële risico's op een fiasco met zich mee. Een eerste onderzoek van SEO Economisch Onderzoek (Van Rij, Ter Beek, Koopmans, 2006) wees uit dat de invoeringskosten hoog zouden zijn en de baten betrekkelijk gering. Uit dat SEO-rapport blijkt dat de (eenmalige) kosten van normalisering tussen de 76 en 245 miljoen euro liggen, terwijl de baten op jaarbasis tussen de 5 en 7,6 miljoen euro bedragen. In het somberste scenario duurt het dan ook bijna een halve eeuw om tot besparingen te komen. 'De terugverdientijd van de aanvankelijke investering varieert tussen tien en vijftig jaar, met 23 jaar als meest waarschijnlijke waarde.' Daarmee leek het argument van voorstanders dat het wetsvoorstel zou bijdragen aan de nodige besparingen niet door het onderzoek ondersteund te worden. Toenmalig minister Ter Horst van BZK baseerde haar afwijzing van een verdere normalisering voor het rijkspersoneel mede op deze conclusies.

H.264/AVC is the newest video coding standard developed by the joint eflort of ITU-T KEG and ISUIIEC MPEG. This standard achieves a significant improvement in coding efficiency relative to former standards at the cost of increased complexiq, thus gaining a lot of attention by industv, but creating a big challenge for efficient hardware and software implementations. In this paper three implementations of an H. 264/A VC baseline decoder using different leading high-end Digitals Signal Processors are described. The implementation has been done, using the JVT reference sofhvare, by undergraduate seniors under the supervision of the authors, in the course of one year. This study is benejkial in several aspects. First, the complexiq of an H.264/AVC decoder is measured and analyzed using "real world" hardware and software. Second, the dtficulfies that programmers could face when implementing an H.264/A VC decoder scheme are examined. And lastly, diflerent Digital Signal Processor...

A pipelined optical-flow processing system that works as a virtual motion sensor has been described. It is based on a field programmable gate array (FPGA) device enabling the easy change of configuring parameters to adapt the sensor to different speeds, light conditions and other environmental factors. It is referred to as a 'virtual sensor' because it consists of a conven- tional camera as front-end supported by an FPGA processing device, which embeds the frame grabber, optical-flow algorithm implementation, output module and some configuration and storage circuitry. This is the first fully stand-alone working optical-flow processing system to include both accuracy and speed of measurement of the platform performance. The customisability of the system for different hardware resources and platforms has also been discussed, showing the resources and performance for a stand-alone board and a PCI co-processing board. Optical flow computation consists in extracting a dense veloc...

The exponential evolution of the smart camera processing performances is directly linked to the improvements on hardware processing elements. Nowadays, high processing performances can be reached considering hardware targets which enables a high level of task parallelism to be implemented. Highly regular tasks are good candidate for a reconfigurable logic implementation and less regular parts of the algorithm could be described on the processor. Meanwhile the prototyping time is related to the selected target and the associated development methodology. The implementation on reconfigurable logic is highly efficient in exploiting the intrinsic task parallelism nevertheless can be time consuming using traditional methodology (i.e. Hardware Language Description). Several approaches can be considered to decrease the proto-typing time and to conserve high processing performances for instance implementation based on: • heterogeneous architectures [1] that mixed reconfig-urable logic (i.e. ...

This papers presents a framework for detecting multiple moving moving objects in a sequence of images. Using a statistical approach, where the inter-frame difference is modeled by a mixture of two Laplacian distributions and a deformable contour-based energy minimization approach, we reformulate the motion detection problem as a front propagation problem. Following the work of geodesic active contours 2], we transform the detection of moving objects problem into an equivalent problem of geodesic computation, which is solved using a level set formulation scheme. To reduce the computational cost required by a direct implementation of the formulation scheme 5] the Narrow Band technique is used. In order to further reduce the CPU time, a multi-scale approach has also been considered. Very promising experimental results are provided using real video sequences.

Without the ground truth image, there is no proper objective method to evaluate sprite generation. In this paper, we propose several complex camera motion patterns to generate synthetic video from original images. Our camera motion patterns include rotation, affine, and Pan-Tilt-Zoom (PTZ) transformations. In addition, our camera motion patterns also include combined patterns. Subsequently, we applied sprite generation to the synthetic videos. Objective evaluation is performed by comparing the ground truth image and sprite based on Peak-Signal-to-Noise Ratio (PSNR) and size. Pattern algebra file is also provided for estimating the accuracy of sprite generation about global motion parameters. Our result indicates that frame PSNR, picture PSNR, size, ground truth image, and pattern algebra file are good indicators of sprite quality.

The diffusion orientation distribution function (ODF) can be reconstructed from q-ball imaging (QBI) to map the complex intravoxel structure of water diffusion. As acquisition time is particularly large for high angular resolution diffusion imaging (HARDI), fast estimation algorithms have recently been proposed, as an on-line feedback on the reconstruction accuracy. Thus the acquisition could be stopped or continued on demand. We adapt these real-time algorithms to the mathematically correct definition of ODF in constant solid angle (CSA), and develop a motion detection algorithm upon this reconstruction. Results of improved fiber crossing detection by CSA ODF are shown, and motion detection was implemented and tested in vivo.

A moving object that has many complex moving parts is very hard to detect and its motion is not easy to estimate. In this paper, we present a new technique for motion estimation and detection of moving complex objects by analyzing the resampled motions of the parts of objects. The Kalman filter is used to track all resampled movements and the tracked routes are classified into groups that share the same fundamental movements. Our simulation show that recall of motion estimation and detection is approximately 0.8, while the computation drops exponentially.

Rate vs. quality is a crucial trade off not only for efficient video coding and transmission but also for adaptive transmission strategies in wireless networks and/or congestion-prone networks. Scalable coders are well suited to tackle the time-varying capacities of these environments. In this paper we propose a semi-analytical model suitable for the medium grain scalable option of the H.264 standard and discuss the parameters influencing its performance. Results show it can effectively be used to represent the expected rate for different quality layers and thus its applicability to algorithms for resource optimization.

This paper addresses the problem of image sequence segmentation. A technique using a sequence model based on compound random fields is presented. This technique is recursive in the sense that frames are processed in the same cadency as they are produced. New regions appearing in the sequence are detected by a morphological procedure.

Variable Block Size (VBS) based motion estimation has been adapted in state of the art video coding, such as H.264/AVC, VC-1. However, a low complexity H.264/AVC encoder cannot take advantage of VBS due to its power consumption requirements. In this paper, we present a VBS partition algorithm based on a binary motion edge map without either initial motion estimation or Rate-Distortion (R-D) optimization for selecting modes. The proposed algorithm uses the Walsh Hadamard Transform (WHT) to create a binary edge map, which provides a computational complexity cost effectiveness compared to other light segmentation methods typically used to detect the required region.

In this paper, we propose a new low complexity video compression method based on detecting blocks containing moving edges using only DCT coefficients. The detection, whilst being very efficient, also allows efficient motion estimation by constraining the search process to moving macro-blocks only. The encoders PSNR is degraded by 2dB compared to H.264/AVC inter for such scenarios, whilst requiring only 5% of the execution time. The computational complexity of our approach is comparable to that of the DISCOVER codec which is the state of the art low complexity distributed video coding. The proposed method finds blocks with moving edge blocks and processes only selected blocks. The approach is particularly suited to surveillance type scenarios with a static camera.