BECAUSE CONVENTIONAL SOFTWARE-BASED PACKET INSPECTION algorithms HAVE NOT KEPT PACE WITH HIGH-SPEED NETWORKS, INTEREST HAS TURNED TO USING HARDWARE TO PROCESS NETWORK DATA QUICKLY. STRING SCANNING WITH BLOOM FILTERS C...
详细信息
BECAUSE CONVENTIONAL SOFTWARE-BASED PACKET INSPECTION algorithms HAVE NOT KEPT PACE WITH HIGH-SPEED NETWORKS, INTEREST HAS TURNED TO USING HARDWARE TO PROCESS NETWORK DATA QUICKLY. STRING SCANNING WITH BLOOM FILTERS CAN SCAN ENTIRE PACKET PAYLOADS FOR PREDEFINED SIGNATURES AT MULTI-GIGABIT-PER-SECOND LINE SPEEDS.
Selective encryption is a technique to save computational complexity or enable interesting new system functionality by only encrypting a portion of a compressed bitstream while still achieving adequate security. Altho...
详细信息
Selective encryption is a technique to save computational complexity or enable interesting new system functionality by only encrypting a portion of a compressed bitstream while still achieving adequate security. Although suggested in a number of specific cases, selective encryption could be much more widely used in consumer electronic applications ranging from mobile multimedia terminals through digital cameras were it subjected to a more thorough security analysis. We describe selective encryption and develop a simple scalar quantizer example to demonstrate the power of the concept, list a number of potential consumer electronics applications, and then describe an appropriate method for developing and analyzing selective encryption for particular compression algorithms. We summarize results from application of this method to the MPEG-2 video compression algorithm.
Image-based rendering (IBR) involves constructing an image from a new viewpoint, using several input images from different viewpoints. A rendering framework must be flexible enough to handle this kind of data. It must...
详细信息
Image-based rendering (IBR) involves constructing an image from a new viewpoint, using several input images from different viewpoints. A rendering framework must be flexible enough to handle this kind of data. It must also be able to handle data extracted from several input images using far less accurate techniques--for example, stereo matching, which introduces a high level of uncertainty. The authors' attempt to define this uncertainty and use it in their rendering framework. When synthesizing new images from real input images, IBR techniques often neglect to take advantage of all the knowledge gathered from computer vision techniques during depth recovery. The authors' image-rendering approach involves three steps: depth extraction, uncertainty estimation, and rendering. Uncertainty is a vague term, which can refer to various aspects of the input images. For example, it can relate to the intrinsic or extrinsic camera parameters. Their technique of rendering depth with uncertainty using ellipsoidal Gaussian kernels is straightforward. They represent each depth uncertainty as an ellipsoid. INSET: Perspective Projection Approximation..
The article describes an approach to explore music collections by navigating through different views. The technique used is called aligned self-organizing maps (SOM). The SOM is an unsupervised neural network with app...
详细信息
The article describes an approach to explore music collections by navigating through different views. The technique used is called aligned self-organizing maps (SOM). The SOM is an unsupervised neural network with applications in various domains including audio analysis. As a clustering algorithm, the SOM is very similar to other partitioning algorithms such as K-means. In terms of topology preservation for visualization of high-dimensional data, alternatives include multi-dimensional scaling, Sammon's mapping and generative topographic mapping. The objective of the SOM is to map high-dimensional data into two-dimensional map in such a way that similar items are located near each other. The SOM consists of an ordered set of units that are arranged in two-dimensional visualization called the map. Common choices to arrange the map units are rectangular or hexagonal grids. There are two main parameters for the SOM algorithm. One is the map size the other is the final neighborhood radius. A larger map gives a higher resolution of the mapping but is computationally more expensive. The final neighborhood radius defines the smoothness of the mapping and should be adjusted depending on the noise level in the data. The SOM is a useful tool for exploring a data set according to a given similarity measure. An aligned SOM is an extension to the basic SOM that allows for interactively shifting the focus between different aspects and exploring the resulting gradual changes in the organization of the data.
作者：
Peserico, EMIT
Comp Sci & Artificial Intelligence Lab Cambridge MA 02139 USA
In the context of competitive analysis of online algorithms for the k-server problem, it has been conjectured that every randomized, memoryless online algorithm exhibits the highest competitive ratio against an offlin...
详细信息
In the context of competitive analysis of online algorithms for the k-server problem, it has been conjectured that every randomized, memoryless online algorithm exhibits the highest competitive ratio against an offline adversary that is lazy, i.e., that will issue requests forcing it to move one of its own servers only when this is strictly necessary to force a move on the part of the online algorithm. We prove that, in general, this lazy adversary conjecture fails. Moreover, it fails in a very strong sense: there are adversaries which perform arbitrarily better than any other adversary which is even slightly "lazier."
暂无评论