Triangle meshes are widely used in representing surfaces in computer vision and computer graphics. Although 2D image processing-based edge detection techniques have been popular in many application areas, they are not...
详细信息
Triangle meshes are widely used in representing surfaces in computer vision and computer graphics. Although 2D image processing-based edge detection techniques have been popular in many application areas, they are not well developed for surfaces represented by triangle meshes. This paper proposes a robust edge detection algorithm for triangle meshes and its applications to surface segmentation and adaptive surface smoothing. The proposed edge detection technique is based on eigen analysis of the surface normal vector field in a geodesic window. To compute the edge strength of a certain vertex, the neighboring vertices in a specified geodesic distance are involved. Edge information are used further to segment the surfaces with the watershed algorithm and to achieve edge-preserved, adaptive surface smoothing. The proposed algorithm is novel in robustly detecting edges on triangle meshes against noise. The 3D watershed algorithm is an extension from previous work. Experimental results on surfaces reconstructed from multi-view real range images are presented.
A novel multilayer neural network (NN) controller in discrete-time is designed to deliver a desired tracking performance for a class of nonlinear systems. A reinforcement learning scheme in discrete-time is proposed f...
详细信息
A novel multilayer neural network (NN) controller in discrete-time is designed to deliver a desired tracking performance for a class of nonlinear systems. A reinforcement learning scheme in discrete-time is proposed for the NN controller, where the learning is performed based on a certain performance measure, which is supplied from a critic. In other words, the critic conveys much less information than the desired output required in supervisory learning. Nevertheless, their ability to generate correct control actions makes adaptive critics prime candidates. The adaptive generating NN in the adaptive critic NN controller approximates the dynamics of the nonlinear system whereas a second NN based critic generates a signal, which is used to tune the weights of the action generating NN. Using the Lyapunov approach, the uniform ultimate boundedness. (UUB) of the closed-loop tracking error and weight estimates are shown by using novel weight updates. The adaptive critic NN does not require an offline learning phase and the weights can be initialized at zero or randomly. Simulation results verify the theoretical conclusions.
The overall quality of network connections has a significant impact on the performance of networked applications. Consequently, QoS management for networked multimedia applications over IP is a significant and immedia...
详细信息
The overall quality of network connections has a significant impact on the performance of networked applications. Consequently, QoS management for networked multimedia applications over IP is a significant and immediate challenge. While differentiated services (diffserv) provide a sense of resource allocation and QoS, they do not guarantee QoS. This paper presents design, implementation and evaluation of a content-aware bandwidth broker (CABB) that manages QoS of multimedia applications for the diffserv environment. CABB allocates network resources to multimedia flows based on client requirement and the tolerance and adaptability of the application. CABB has been implemented and evaluated using the NS-2 simulator toolkit. Evaluations show that CABB improves network resource allocations and increases overall throughput. The results also show that multimedia flows are better managed and controlled, thereby improving perceived QoS and avoiding possible congestion in core routers.
This paper describes the performance benefits attained using enhanced network interfaces to achieve low latency communication. We make use of DMA communication mode, to send data to other nodes, while the CPU performs...
详细信息
ISBN:
(纸本)9780769515823
This paper describes the performance benefits attained using enhanced network interfaces to achieve low latency communication. We make use of DMA communication mode, to send data to other nodes, while the CPU performs useful calculations. Zero-copy communication is achieved through pinned-down physical memory regions, provided by NIC's driver modules. Our testbed concerns the parallel execution of tiled nested loops onto a Linux PC cluster with PCI-SCI NICs (Dolphin D330). Tiles are essentially exchanging data and should also have large Computational grain, so that their parallel execution becomes beneficial. We schedule tiles much more efficiently by exploiting the inherent overlapping between communication and computation phases among successive, atomic tile executions. The applied nonblocking schedule resembles a pipelined data-path where computation phases are overlapped with communication ones, instead of being interleaved with them. Experimental evaluation illustrates that when using enhanced communication features such as DMA transfers, memory-mapped interfaces and zero-copy mechanisms, overall performance is considerably improved compared to using conventional, CPU and kernel bounded, communication primitives.
Limited authority of actuators implies that real control signals are always constrained, and in almost all cases, this produces a degradation in the performance of the system. It is thus of practical importance to und...
详细信息
Limited authority of actuators implies that real control signals are always constrained, and in almost all cases, this produces a degradation in the performance of the system. It is thus of practical importance to understand the fundamental aspects of this performance degradation. In this paper, we take an initial step by proposing a way to characterize the performance limitations that arise in closed-loop stable linear systems due to the constraints on the magnitude of the control signal. Specifically, we evaluate the cost associated with the constraints via the L 2 norm of the tracking error of a constrained limiting optimal compensator.
A novel real-time discrete-event admission control (AC) scheme for high-speed networks is proposed with the aim of attaining a desired quality of service (QoS) and high network utilization. The AC encapsulates an adap...
详细信息
A novel real-time discrete-event admission control (AC) scheme for high-speed networks is proposed with the aim of attaining a desired quality of service (QoS) and high network utilization. The AC encapsulates an adaptive bandwidth estimation scheme, a congestion indicator, and a peak bit/cell rate (PBR/PCR) estimate from new sources. The admission controller dynamics are expressed as a discrete-event system while the bandwidth estimator design is based on Lyapunov theory and the congestion controller dynamics are given in discrete-time. The novel aspect of the proposed approach is the application of hybrid system theory to prove the performance of the proposed admission controller, stability and the development of a rigorous and repeatable design procedure. The performance of the proposed AC is evaluated using the QoS metrics, which are given in terms of service delay, packet/cell losses, and network utilization. Simulation results are presented by streaming ON/OFF and MPEG video data into the network. Results show that the proposed AC admits significantly more traffic compared to other available admission control schemes thereby guaranteeing high network utilization while maintaining the desired QoS.
Many real-time systems employed in defense, space, and consumer applications have power constraints and high reliability requirements. In this paper, we focus on the relationship between fault tolerance techniques and...
详细信息
Many real-time systems employed in defense, space, and consumer applications have power constraints and high reliability requirements. In this paper, we focus on the relationship between fault tolerance techniques and energy consumption. In particular, we establish the energy efficiency of Application Level Fault Tolerance (ALFT) over other software-based fault tolerance methods. We then develop sensible energy-aware heuristics for ALFT schemes. The heuristics yield up to 40% energy savings.
This paper proposes an adaptive methodology to prevent congestion in packet switched networks such as the internet, where the internal network nodes convey very little information to the ingress nodes. A novel archite...
详细信息
This paper proposes an adaptive methodology to prevent congestion in packet switched networks such as the internet, where the internal network nodes convey very little information to the ingress nodes. A novel architecture of preventing the congestion is presented when the traffic arrival rates and bottleneck queue levels are known. In the architecture, the congestion is controlled by adjusting the transmission rates of non real-time and certain real-time sources in response to the feedback information so that a desired quality of service (QoS) can be met instead of using the existing additive increase multiplicative decrease (AIMD) algorithm. The QoS is defined in terms of packet loss, transmission delay, network utilization and fairness. Mathematical analysis is given to demonstrate the stability of the closed-loop system. The proposed methodology can be readily applied to network planning, designing routing algorithms and transmission links as well as for real-time video and voice data transfer in unicast networks.
Cross-correlation based time delay estimates (TDE) can be used for direction-of-arrival (DOA) estimation with an acoustic array in not-too-reverberant environments. In order to benefit from the computational efficienc...
详细信息
Cross-correlation based time delay estimates (TDE) can be used for direction-of-arrival (DOA) estimation with an acoustic array in not-too-reverberant environments. In order to benefit from the computational efficiency of TDE-based DOA estimation, and concentrating on applications that use a compact microphone array and low sampling frequency, we use a combination of approaches to make TDE-based DOA estimation more robust under reverberant conditions. Cross-correlation interpolation based on the Goertzel algorithm is used for generalized cross-correlation TDE as well as in the steered-response-power algorithm used for comparison. The generalized cross-correlation TDE algorithm precedes a robust TDE-to-DOA process. The latter includes DOA and crosscorrelation dependent weighting of the various TDE and removal of an outlier TDE. When the use of TDE from consecutive frames is possible, further performance improvement results from unit-norm DOA adaptation. Performance is evaluated for simulated and recorded data.
暂无评论