Abstract

Abstract

Abstract
This paper presents a new approach to synthesize to reconfigurable hardware (HW) user-specified regions of a program, under the assumption of "virtual HW" support. The automation of this approach is supported by a compiler front-end and by an HW compiler under development. The front-end starts from the Java bytecodes and, therefore, supports any language that can be compiled to the JVM (Java Virtual Machine) model. It extracts from the bytecodes all the dependencies inside and between basic blocks. This information is stored in representation graphs more suitable to efficiently exploit the existent parallelism in the program than those typically used in high-level synthesis. From the intermediate representations the HW compiler exploits the temporal partitions at the behavior level, resolves memory access conflicts, and generates the VHDL descriptions at register-transfer level that will be mapped into the reconfigurable HW devices.

Abstract
In this article, two artificial neural networks based models for motion detection and motion analysis in image sequence, are proposed and analyzed. The first model is applicable to sequences acquired by a static camera, and segments the images into two areas: moving objects and background. The second model, applicable to image sequences obtained by a moving camera, uses a neural network to identify the `interesting points' in the images (e.g. edges, corners), and a block matching algorithm to calculate the movement in these points. Several tests with real images have been performed to evaluate the models. The results are similar to those obtained by other methods but the proposed techniques show some advantages which are discussed in the paper.

Abstract
The risk of the duration of a project exceeding a certain limit has been studied since the early papers proposing the PERT method but no results are available relating the statistical distribution of the total duration with the morphological features of the network. In this paper, a set of indicators is proposed to describe the morphology of the project network and an experimental research is carried out to relate the distribution of the total duration with such indicators. A random generator of project networks is used as well as a model to produce graphical representations of the network. Major results are presented showing how sensitive are the parameters of the distribution of the total duration to the morphology of the network, allowing the project manager to estimate the risk of delay in terms of the proposed indicators. Thus, these results can be used as a decision aid to select alternative designs for the project network taking into account the correspondent risk of delay.

Abstract