Catálogo de publicaciones - libros

In this study the effects of using VPNs on the throughput of wireless LANs are investigated. To this end we have set up a test bed and based our observations on the actual measurements obtained from the experiments performed using this test bed. The wireless LAN we have used is an IEEE 802.11g network and we compared the throughput values obtained by employing IPSec in this network with the calculated ones. We performed the tests using periodically generated traffic with 1ms and 5 ms intervals between successive packets. We have also considered the case for random traffic generated according to exponential distribution. We present the variations of the throughput with respect to packet size for different input traffic patterns.

In this paper a novel optical burst assembly technique, namely path loss rate driven burst assembly which is an enhancement on an adaptive threshold based burst assembly is proposed. Path loss rate driven burst assembly adjusts the threshold values based on the feedback obtained from the network. Each egress node sends feedback packets periodically to each ingress node informing them on the contention level along the path. As the traffic along the corresponding path becomes lighter, an ingress node allows the corresponding burst generator to use greater time and size threshold values. Conversely, under heavy traffic along the path, the corresponding burst generators are forced to decrease the time and size threshold parameters. The results are obtained in terms of byte loss rate and delay, and compared with those obtained under nonadaptive (static) burst assembly and previously proposed link loss rate based adaptive threshold burst assembly. It is observed that the proposed technique significantly decreases the byte loss rate as the traffic gets heavier by keeping the end to end delay in a feasible range.

A new pixel domain spatio-temporal video noise filter for archive film restoration has been proposed in this paper. The proposed filtering method takes motion changes and spatial information into account. Firstly, temporal filtering is carried out considering temporal changes adaptively. Afterwards, interpolation between degraded and temporally filtered images is carried out to preserve edge information using local standard deviation values. With respect to pixel domain techniques proposed in the literature, the proposed method gives better results for various test videos and particularly provides superior results for archive film.

In this study sub-pixel motion estimation using one-bit transform is presented. The utilization of low bit depth representations for full-pixel motion estimation exist in the literature [1,7]. However sub-pixel motion estimation employing these kinds of transforms has not been tried until now. In this paper, it is shown that one-bit transform can also be utilized for sub-pixel motion estimation, significantly improving motion estimation accuracy. Hence a fast sub-pixel motion estimation approach is facilitated using one-bit transform.

In this paper, we present a new technique to find the optimal neighborhood decomposition for convex structuring elements used in morphological image processing on hexagonal grid. In neighborhood decomposition, a structuring element is decomposed into a set of neighborhood structuring elements, each of which consists of the combination of the origin pixel and its six neighbor pixels. Generally, neighborhood decomposition reduces the amount of computation required to perform morphological operations such as dilation and erosion. Firstly, we define a convex structuring element on a hexagonal grid and formulate the necessary and sufficient condition to decompose a convex structuring element into the set of basis convex structuring elements. Secondly, decomposability of a convex structuring element into the set of primal bases is also proved. Furthermore, cost function is used to represent the amount of computation or execution time required for performing dilations on different computing environments and by different implementation methods. The decomposition condition and the cost function are applied to find the optimal neighborhood decomposition of a convex structuring element, which guarantees the minimal amount of computation for morphological operations. Example decompositions show that the decomposition results in great reduction in the amount of computation for morphological operations.

Modified two-bit transform based block motion estimation is presented in this paper. Initially video frames are converted into two-bit representations using the two-bit transform (2BT) and binary block based motion estimation is performed using these two bit-planes. Modification to the original 2BT based motion estimation scheme is introduced by conditional local or full searches using the MAD criterion to improve the initial motion estimation accuracy. Experimental results show that the proposed modified 2BT-based motion estimation technique can significantly improve peak signal-to-noise radio performance compared to 2BT without modification, and also outperforms the modified one-bit transform (1BT) based motion estimation approach.

Recently, researchers proposed deterministic and statistical appearance-based 3D face tracking methods that can successfully tackle the image variability and drift problems. However, appearance-based methods dedicated to 3D face tracking may suffer from inaccuracies since they are not very sensitive to out-of-plane motion variations. On the other hand, the use of dense 3D facial data provided by a stereo rig or a range sensor can provide very accurate 3D face motions/poses. However, this paradigm requires either an accurate facial feature extraction or a computationally expensive registration technique (e.g., the Iterative Closest Point algorithm). In this paper, we propose a 3D face tracker that is based on appearance registration and on a fast variant of a robust Iterative Closest Point algorithm. The resulting 3D face tracker combines the advantages of both appearance-based trackers and 3D data-based trackers. Experiments on real video data show the feasibility and usefulness of the proposed approach.

This paper presents an improved cross-correlation algorithm for template-based object tracking: the reduced spiral search with a sparse angular sampling. The basic parameters of the algorithm for the real-time face tracking were evaluated regarding their impact on the algorithm performance. They are the minimum number of pixels and the size of the template, the correlation threshold and drifting, and the parameters of the search – radius, shift, direction, and rotation of the template. We demonstrated that the information provided by the grid-like template might be reduced to 16 pixels with a grid step of 15 pixels. A spiral search in 8 directions with a minimum shift of 1 pixel decreases the number of computations by 20 times. Being activated sequentially the template rotation does not increase the performance, but doing the tracking adaptive and robust.

Locating human fiducial points like eyes and mouth in a frontal head and shoulder image is an active research area for applications such as model based teleconferencing systems, model based low bit rate video transmission, computer based identification and recognition systems. This paper proposes an adept and efficient rule based skin color region extraction algorithm using normalized r-g color space. The given scheme extracts the skin pixels employing a simple quadratic polynomial model and some additional color based rules to extract possible eye and lip regions. The algorithm refines the search for fiducial points by eliminating falsely extracted feature components using spatial and geometrical representations of facial components. The algorithm described herein has been implemented and tested with 311 images from FERET database with varying light conditions, skin colors, orientation and tilts. Experimental results indicate that the proposed algorithm is quite robust and leads to good facial feature extraction.

Medical imaging applications produce large sets of similar images. Thus a compression technique is useful to reduce space storage. Lossless compression methods are necessary in such critical applications. Volumetric medical data presents strong similarity between successive frames. In this paper we investigate predictive techniques for lossless compression of video sequences applied to volumetric data. We also make a comparative study with other existing compression techniques dedicated to volumetric data.