首页> 外文会议>43rd annual Midwest Instruction and Computing Symposium 2010 >A Concurrent Image Filtering Project for Undergraduate Computer Science Education
【24h】

A Concurrent Image Filtering Project for Undergraduate Computer Science Education

机译:本科计算机科学教育并行图像过滤项目

获取原文
获取原文并翻译 | 示例

摘要

Parallel and distributed computing has emerged as an increasingly important topic in Computer Science education driven largely by the predominance of multi-core systems, cloud computing systems and the use of GPU's for general computation. As fundamental concepts of concurrency become more tightly integrated with undergraduate education there is a heightened need for projects that have both a practical and accessible parallel solution. This paper presents a parallel programming project using Java that is based on median filtering digital images. The project is designed for use in an undergraduate computer science curricula and has been used in a programming languages course. The project has appealing features since 1) the problem domain lies within the field of multimedia and digital image processing 2) the problem domain is easily understood 3) the sequential solution is straightforward to implement 4) the output can be visualized since the output is a digital image and 5) parallelization provides a clear speedup where the controlling parameters can be tuned to exhibit both the cost and benefit of synchronization. Image processing is a field of study within the broader domain of signal processing where a digital image, commonly referred to as the source image, is processed to produce an output digital image, which is commonly referred to as the destination image. Digital images are often filtered to produce an image that is more suitable for later computation or to produce an image that is more aesthetically or artistically pleasing than the source image. Median filtering is a non-linear process based on a straightforward statistical analysis of a local sub-region of an image. This paper gives an overview of median filtering digital images, provides a naive sequential solution, a design for parallelization, a discussion of the performance gains and concludes with a discussion of issues related to using the project in an undergraduate educational setting.
机译:在计算机科学教育中,并行和分布式计算已成为一个越来越重要的话题,这主要是由多核系统,云计算系统的优势以及将GPU用于常规计算的推动所致。随着并发基本概念与本科教育的紧密结合,对具有实用且可访问的并行解决方案的项目的需求日益增加。本文介绍了一个基于Java的并行编程项目,该项目基于中值滤波数字图像。该项目旨在用于本科计算机科学课程,并已用于编程语言课程。该项目具有吸引人的特征,因为1)问题域位于多媒体和数字图像处理领域2)问题域易于理解3)顺序解决方案易于实现4)输出可以可视化,因为输出是数字图像和5)并行化提供了明显的加速,可以调整控制参数以显示同步的成本和收益。图像处理是信号处理更广泛领域中的一个研究领域,其中,通常将数字图像(通常称为源图像)进行处理以产生输出数字图像(通常将其称为目标图像)。通常对数字图像进行滤波,以生成更适合于以后计算的图像,或者生成比源图像更具美学或艺术美感的图像。中值滤波是基于图像局部子区域的直接统计分析的非线性过程。本文概述了中值滤波数字图像,提供了一个简单的顺序解决方案,一个并行化设计,一个性能提升的讨论,并以与在本科教育环境中使用该项目有关的问题作为结尾进行了讨论。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号