Maria Stela Veludo de Paiva

Comparative study of the server-initiated lowest algorithm using a load balancing index based on the process behavior for heterogeneous environment

The availability of low cost microcomputers and the evolution of computer networks have increased the development of distributed systems. In order to get a better process allocation on distributed environments, several load balancing algorithms have been proposed. Generally, these algorithms consider as the information policy’s load index the length of the CPU’s process waiting queue. This paper modifies the Server-Initiated Lowest algorithm by using a load index based on the resource occupation. Using this load index the Server-Initiated Lowest algorithm is compared to the Stable symmetrically initiated, which nowadays is defined as the best choice. The comparisons are made by using simulations. The simulations showed that the modified Server-Initiated Lowest algorithm had better results than the Symmetrically Initiated one.

Analysis on the significant information to update the tables on occupation of resources by using a peer-to-peer protocol

The adequate occupation of the computing resources can influence, in a decisive way, the global performance of the system. Therefore, in order to achieve a high performance, it is mandatory to know all the computing resources involved and their respective occupation level in a certain moment. With the objective of improving the system performance, the paper presents the OpenTella model to update the information related to the occupation of resources and the respective analysis of this occupation so that the migration of processes among computers of a same cluster can be completed. With the objective of increasing the scale level in the system and decreasing the number of messages among the computers, this peer-to-peer protocol defines sub-nets, which are clusters that make up a more comprehensive cluster. Thus, groups are defined to interchange information and update the occupation of resources, in order to minimize the communication and to achieve a calculation to balance the load and meet the system needs, resulting in the migration of processes.

OpenTella: A Peer-to-Peer Protocol for the Load Balancing in a System Formed by a Cluster from Clusters

The search for the high performance in the traditional computing systems has been related to the development of specific applications executed over parallel machines. However, the cost of this kind of system, which is high due to the hardware involved in these projects, has limited the continuing development in these areas, as just a small part of the community has access to those systems. With the aim of using a low cost hardware to achieve a high performance, this paper presents the OpenTella, a protocol based on the peer-to-peer models to update the information related to the occupation of resources and an analysis of this occupation for a post migration of processes among computers of a cluster.

True-Motion Estimation and Compensation with Multitemporal Block-Matching Search

in de-interlacing algorithms are fundamental to explore image quality potential on modern TV screen technologies. This paper presents a true motion vector verification algorithm based on multi temporal block matching strategy, applied to video de-interlacing. During field scanning, a block of pixels is selected, then a block with the same reference coordinates is extracted but in the future field. A searching process runs in the previous field, looking for the most similar block; the best fit block and their coordinates are extracted from the previous field, and a new searching process is done but in the future field. If the searching process in the future field results back to the original block coordinates, the motion vector is validated; in other case, a correction over the motion vectors is done. Error calculation showed that the proposed algorithm presents image quality improvement, when compared to a classical motion compensation algorithm.

A Java cluster management service

This paper presents JCM (Java cluster management) project which provides a performance, tool-set and viability analysis of the Java platform to design high-performance applications. This is done through the design of a Java cluster management service based on the evolution of the Java platform, using MAJC (Microprocessor Architecture for Java Computing). The design of the cluster was proposed using queuing theory and UML. The JCM makes a comparison to existent clusters like Beowulf and DSP systems to the cluster management service developed.

High-boost Weber local filter for precise eye localization under uncontrolled scenarios

Abstract The task of precise eye localization for a given face is very important for many computer vision applications such as face recognition, face and eye tracking, and face alignment. Further, this is a challenging task when performed in uncontrolled scenarios. This paper proposes a new method for precise eye localization in uncontrolled scenarios using the high-boost Weber local filter (HBWLF); the proposed filter emphasizes high-frequency components, without eliminating the low-frequency ones, and thus enhances the details of the periocular region (region around the eyes), which is critical to the task of eye localization. The proposed method was evaluated using Labeled Faces in the Wild (LFW) and BioID databases, and the results show that HBWLF yields better performance than several classical methods.

A topological approach for detection of chessboard patterns for camera calibration

Abstract This work aims to automatically identify chessboard patterns for camera calibration. The method uses a fast x-shaped corner detector and a geometric mesh to represent the relative association between features. The mesh allows considering the regularity of the chessboard pattern and a topological filter is presented. The matching between real world points and their image projections is done using neighboring properties in a filtered mesh. The point location is locally updated to the subpixel precision with a specific x-corner detector. The calibration points are determined even when the pattern is partially occluded. The experiments show that the proposed algorithm provides a robust detection of the chessboard patterns and take great advantage of image frames. The method is applicable for both online and off-line detection of chessboard patterns.

Super Resolution Image Created Automatically from Mosaics

In order to extract information from the images, both the resolution and its application must be observed. This resolution depends basically on the physical characteristics of the sensor responsible for capturing these images, i.e. the disposition of the lenses, radiometric resolution, positioning, dimensions of the same, which determines the pixel size acquired by the device, among other technical factors. The objective of this work is to develop efficient algorithms, robust and automated fusion image frames for mosaics and super resolution from the selection of a region in the mosaic. Image registration is a fundamental step in combining several images that make up the scene. Our research is based on the determination and extraction of characteristics defined by the SIFT algorithm (detector / descriptor) and RANSAC. Validations real data show the effectiveness of our techniques proposed. The difference of this work is the way to get the matching and merging of images because it occurs dynamically between elements common images that are stored in a sparse matrix.

B-spline generation in FPGA

B-spline curves can represent almost any smooth shape, being adopted by file exchange standards, used as a wavelet basis for multiresolution systems and manufacturing efficiency is increased using b-splines toolpaths. The respective algorithms for conventional CPUs demand great computational effort, requiring another hardware architectures, like GPUs, for better processing efficiency. This paper describes a system on a chip that generates b-splines curves from given control points in a concurrent way, based in a FPGA platform, including features that are not addressed by the few works in the area. This system is faster than a quadcore CPU in the order of the number of generated points, at the same clock rate. The independence of the number of generated and initial control points, and the processing determined by the FPGA capacity, makes it a better choice for embedded systems in relation to conventional CPUs.

Strategies to run an ophthalmological CW Self-Raman laser in micro-second pulsed laser regime

Micro-second pulsed laser exposure is a new procedure for retina disease treatments. This technique consists in applying laser power sequences of 200µs pulses (1000 times less than traditional treatments). Short laser exposure duration avoids photoreceptor lesions on the retina, providing a better final vision [1]. The laser cavity in 586nm and 4W CW output power is based on Self-Raman conversion in Nd:GdVO4 [2],[3]. Initially, it was developed to continuous mode, but now it is switched to 500Hz, with minimum pulse duration of up to 100µs. This regime is called fast pulse mode. This operation mode makes the cavity PID control system velocity vital and challenging [4]. The energy delivery should be very fast for the pulse to reach the desired level before its termination, and the control feedback system has to correct any deviation of output power to keep it stable and constant. The strategy found was to build a duplicated current driver, one plugged to the laser head and the second plugged to a fake laser cavity. Each one works in synchronism to keep the drained current constant while the laser pulse in turned on and off during treatment procedure. Consequently, it eliminates the power supply time response dependency which can take around 1ms. Fig. 1 shows how the current is switched between both loads against time, fake laser current (I2) and pump diode laser current (I1), and constant drained current from power supply. The output power control is performed by a high priority software to minimize any delay in power reading. The laser head showed good performance, without any stability issues. The cavity pulse time response was of 30µs (worst case) to reach a stable mode. Fig. 2 shows the control loop effort to make output power stable by changing laser diode current. It takes 50ms to achieve the regime state. Thermal lens effects are also investigated. It causes performance differences between CW and micro-pulse mode operation, and laser cavity alignment pulse regime is responsible to favor one mode. This work resulted in a commercial ophthalmic laser.