Resumo em Inglês:

In this paper we describe an aspect-oriented infrastructure to handle dynamic AOP based on the Lua language. This infrastructure is composed of AspectLua, a Lua extension that allows the declaration of aspects, and a meta-object protocol, LuaMOP, that unifies the introspective and reflective mechanisms provided by Lua. Aspects are defined in isolation using AspectLua and then they are weaved through LuaMOP. An important feature of AspectLua is to allow the association of aspects with undeclared elements of the application code (anticipated join points). Furthermore, it combines a range of features to make AOP easier and powerful.

Resumo em Inglês:

The design and development of distributed software is a complex task. This was not different in OurGrid, a project whose objective was to develop a free-to-join grid. After two years of development, it was necessary to redesign OurGrid in order to cope with the integration problems that emerged. This paper reports our experience in using Aspect-Oriented Programming (AOP) in the process of redesigning the OurGrid middleware. The essential direction of our approach was to get the project (and the software) back in shape. We discuss how the lack of separation of concerns created difficulties in the project design and development and how AOP has been introduced to overcome these problems. In particular, we present the event-based pattern designed to better isolate the middleware concerns and the threads. Besides, we also present the aspects designed for managing the threads and for aiding the testing of multithreaded code. We also highlight the lessons learned in the process of regaining control of the software.

Resumo em Inglês:

Metric Access Methods (MAM) are employed to accelerate the processing of similarity queries, such as the range and the k-nearest neighbor queries. Current methods, such as the Slim-tree and the M-tree, improve the query performance minimizing the number of disk accesses, keeping a constant height of the structures stored on disks (height-balanced trees). However, the overlapping between their nodes has a very high influence on their performance. This paper presents a new dynamic MAM called the DBM-tree (Density-Based Metric tree), which can minimize the overlap between high-density nodes by relaxing the height-balancing of the structure. Thus, the height of the tree is larger in denser regions, in order to keep a tradeoff between breadth-searching and depth-searching. An underpinning for cost estimation on tree structures is their height, so we show a non-height dependable cost model that can be applied for DBM-tree. Moreover, an optimization algorithm called Shrink is also presented, which improves the performance of an already built DBM-tree by reorganizing the elements among their nodes. Experiments performed over both synthetic and real world datasets showed that the DBM-tree is, in average, 50% faster than traditional MAM and reduces the number of distance calculations by up to 72% and disk accesses by up to 66%. After performing the Shrink algorithm, the performance improves up to 40% regarding the number of disk accesses for range and k-nearest neighbor queries. In addition, the DBM-tree scales up well, exhibiting linear performance with growing number of elements in the database.

Resumo em Inglês:

This paper proposes a method for color gradient computation applied to morphological segmentation of color images. The weighted gradient (with weights estimated automatically), proposed in this paper, applied in conjunction with the watershed from markers technique, provides excelent segmentation results, according to a subjective visual criterion. The weighted gradient is computed by linear combination of the gradients from each band of an image under the IHS color space model. The weights to each gradient are estimated by a systematic method that computes the similarity between the image to compute the gradient and an "ideal image", whose histogram has an uniform distribution. Several experiments were done in order to compare the segmentation results provided by the weighted gradient to the results provided by other color space metrics, also according to a subjective criterion, and such comparison is present in this paper.

Resumo em Inglês:

This paper presents a non-iterative regularized inverse solution to the image interpolation problem. This solution is based on the segmentation of the image to be interpolated into overlapping blocks and the interpolation of each block, separately. The purpose of the overlapping blocks is to avoid edge effects. A global regularization parameter is used in interpolating each block. In this solution, a single matrix inversion process of moderate dimensions is required in the whole interpolation process. Thus, it avoids the large computational cost due to the matrices of large dimensions involved in the interpolation process. The performance of this approach is compared to the standard iterative regularized interpolation scheme and to polynomial based interpolation schemes such as the bicubic and cubic spline techniques. A comparison of the suggested approach with some algorithms implemented in the commercial ACDSee software has been performend in the paper. The obtained results reveal that the suggested solution has a better performance as compared to other algorithms from the MSE and the edges preservation points of view. Its computation time is relatively large as compared to traditional algorithms but this is acceptable when image quality is the main concern.