Davide Quaglia

New sorting-based lossless motion estimation algorithms and a partial distortion elimination performance analysis

In video encoding, block motion estimation represents a CPU-intensive task. For this reason, many fast algorithms have been developed to improve searching and matching phases. A milestone within the lossless approach is partial distortion elimination (PDE/SpiralPDE) in which distortion is the difference between the block to be coded and the candidate prediction block. In this paper, (i) we analyze distortion behavior from local information using the Taylor series expansion and show that our general analysis includes other previous similar approaches. (ii) Then, we propose two full-search (lossless), fast-matching, block motion estimation algorithms, based on the PDE idea. The proposed algorithms, called fast full search with sorting by distortion (FFSSD) and fast full search with sorting by gradient (FFSSG), sort the contributions to distortion and the gradient values, respectively, in order to quickly discard invalid blocks. Experimental results show that the proposed algorithms outperform other existing full search algorithms, reducing by up to 20% the total CPU encoding time (with respect to SpiralPDE), while the computation strictly required by the motion estimation is reduced by about 30%. (iii) Finally, we experimentally find an operational lower bound (based on standard test sequences) for the average number of checked pixels in the PDE approach, which measures the performance of the searching and matching phases. In particular, SpiralPDE achieves performances very close to the searching phase bound, while there is still a remarkable margin on the matching phase. We then show that our algorithms, aimed at improving the performances of the matching phase, achieve interesting results, significantly approaching this margin.

A SystemC-based framework for modeling and simulation of networked embedded systems

Next-generation networked embedded systems pose new challenges in the design and simulation domains. System design choices may affect the network behavior and network design choices may impact on the system design. For this reason, it is important -at the early stages of the design flow- to model and simulate not only the system under design, but also the heterogeneous networked environment in which it operates. For this purpose, we have exploited a modeling language traditionally used for System design -SystemC- to build a system/network simulator named SystemC Network Simulation Library (SCNSL). This library allows to model network scenarios in which different kinds of nodes, or nodes described at different abstraction levels, interact together. The use of SystemC as unique tool has the advantage that HW, SW, and network can be jointly designed, validated and refined. As a case study, the proposed tool has been used to simulate a sensor network application and it has been compared with NS-2, a well-known network simulator; SCNSL shows nearly two-order-magnitude speed up with TLM modeling and about the same performance as NS-2 with a mixed TLM/RTL scenario. The simulator is partially available to the community at http://sourceforge.net/projects/scnsl/.

Model-based distortion estimation for perceptual classification of video packets

In video communications over IP networks, quality of service (QoS) guarantees must be introduced to limit the effect of packet losses. In particular, end-to-end QoS can be improved if packets are protected according to the distortion that would be introduced at the receiver by their loss. In the traditional analysis-by-synthesis (AbS) approach, each packet is assumed lost, error concealment applied, the sequence decoded, and the resulting overall distortion computed. This process produces reliable distortion estimates, but is computationally demanding. In this work, we present a hybrid approach: the distortion introduced in the current frame is evaluated with the AbS method, while the distortion in future frames is estimated by means of a statistical error-propagation model. Results obtained on eight, widely different H.264 sequences show that the proposed model successfully estimates overall distortion with very low complexity. Network simulations also show that model-based packet classification, when used for video transmission over DiffServ networks, delivers PSNR results which are consistently within 0.1 dB compared to the AbS technique.

The COMPLEX reference framework for HW/SW co-design and power management supporting platform-based design-space exploration

The consideration of an embedded devices power consumption and its management is increasingly important nowadays. Currently, it is not easily possible to integrate power information already during the platform exploration phase. In this paper, we discuss the design challenges of todays heterogeneous HW/SW systems regarding power and complexity, both for platform vendors as well as system integrators. As a result, we propose a reference framework and design flow concept that combines system-level power optimization techniques with platform-based rapid prototyping. Virtual executable prototypes are generated from MARTE/UML and functional C/C++ descriptions, which then allows to study different platforms, mapping alternatives, and power management strategies. Our proposed flow combines system-level timing and power estimation techniques available in commercial tools with platform-based rapid prototyping. We propose an efficient code annotation technique for timing and power properties enabling fast host execution as well as adaptive collection of power traces. Combined with a flexible design-space exploration (DSE) approach our flow allows a trade-off analysis between different platforms, mapping alternatives, and optimization techniques, based on domain-specific workload scenarios. The proposed framework and design flow has been implemented in the COMPLEX FP7 European integrated project.

System/network design-space exploration based on TLM for networked embedded systems

This article presents a methodology for the design of Networked Embedded Systems (NESs), which extends Transaction Level Modeling (TLM) to perform system/network design-space exploration. As a result, a new design dimension is added to the traditional TLM refinement process to represent network configuration alternatives. Each network configuration can be used to drive both architecture exploration and system validation after each refinement step. A system/network simulation taxonomy is investigated aiming at precisely identifying the role of cosimulation in system/network design-space exploration. Furthermore, a general criterion to map functionalities to system and network models is presented. As a case study, the proposed methodology is applied to the design of a Voice-over-IP client.

Distortion-based packet marking for mpeg video transmission over diffserv networks

We present a distortion-based approach to packet classification for multimedia transmission over differentiatedservices packet networks. Instead of sending all traffic as premium or relying on a priori data partitioning, packets are individually examined and assigned to different service classes depending on the level of distortion that their loss would introduce at the decoder. Applied to video sequences encoded with the ISO MPEG-2 video coding standard, the proposed distortion-based packet marking scheme outperforms source-transparent techniques and provides substantial and consistent gains in PSNR over the regular best-effort case sending as little as 10% of the packets as premium traffic. Video samples are available at

Perceptually optimized MPEG compression of synthetic video sequences

This paper addresses the problem of improving the quality performance of synthetic video sequences by means of standard frame-based coders. The proposed technique can exploit both the knowledge of the 3D model and the intermediate information computed during the rendering process. Firstly, objects are classified, either semantically or automatically, according to their importance. Then the object classification is translated into a macroblock classification, with particular attention to object boundaries. The classification influences the encoder parameters selection, for instance, the quantization parameter. In order to maximize the performance, we propose a rate-distortion formulation of the problem. Experimental results compared with model-unaware encoding show that the proposed techniques can deliver consistent visual quality improvements for different synthetic scenarios using the same bitrate or even less. Demo sequences are available at http://media.polito.it/perceptual3d.

Delivery of MPEG video streams with constant perceptual quality of service

Constant levels of perceptual quality of service is what ideally users of multimedia services expect. In most cases, however, they receive time-varying levels of quality of service. This paper describes a technique to deliver nearly constant perceptual quality of service when transmitting video sequences over differentiated services IP networks. MPEG video packets are transmitted either as low-loss premium packets or as regular best-effort packets depending on their individual perceptual importance. On a frame-by-frame basis, allocation to the premium class is performed depending on the perceptual importance of each macroblock, the desired level of quality of service and the instantaneous network state. The resulting perceptually-based, time-varying use of premium and best-effort network resources delivers nearly constant quality of service to end users and it yields significant higher PSNR values compared to constant allocation of premium bandwidth.

Improving Performance of Networked Control Systems by Using Adaptive Buffering

The performance of networked control systems is strongly affected by time-varying transmission delays. A traditional solution to this problem consists of storing arriving packets in a buffer which smooths delay jitter at the cost of an increased constant delay. The size of the buffer is based on either a long-term or worst case analysis of network behavior leading to poor performance when the instantaneous network behavior is different. To overcome this problem, this paper proposes the following: 1) to adapt the buffer size according to the actual delay variation; 2) to resize buffer content by using cubic spline smoothing which also reduces the signal noise; and 3) to use a Smith predictor at the controller side. Simulation results show that the adaptive buffering strategy reduces delay and packet loss probability while the spline smoothing process improves control performance even in case of constant-size buffers.

A SystemC/Matlab co-simulation tool for networked control systems

Abstract Real-time systems connected through packet networks belong to the family of networked control systems, and they can be easily destabilized by communication delay and packet losses, when they are not properly compensated. The largest part of the solutions available in the literature are mainly based on control and system theory where the parameters of the network are assumed to be given. This classical approach could be improved by designing at the same time the network, e.g., by introducing quality-of-service guarantees as currently done in teleconference applications. Such control/network co-design needs a simulation framework where both aspects are properly and jointly addressed. The paper addresses this topic starting from the discussion of its critical issues, and then proposing an accurate co-simulation tool based on SystemC and Matlab/Simulink. SystemC will be used for the network simulation and protocol design whereas Matlab/Simulink for plant modeling and control design.