首页    期刊浏览 2025年12月03日 星期三
登录注册

文章基本信息

  • 标题:MISD Compiler for Feature Vector Computation in Serial Input Images
  • 本地全文:下载
  • 作者:Lucas Leiva ; Nelson Acosta
  • 期刊名称:ARPN Journal of Systems and Software
  • 电子版ISSN:2222-9833
  • 出版年度:2011
  • 卷号:1
  • 期号:3
  • 页码:108-116
  • 出版社:ARPN Publishers
  • 摘要:

    In this paper a compiler capable of generate Multiple Instruction Single Data (MISD) architectures for feature vector calculation is presented. The input is a high-level language, avoiding to developers to involve in low level design. Instead, the output is expressed in a Hardware Description Language (HDL), and can be used for FPGA configuration. A FPGA is a programmable device which allows parallelism, increasing the system speed up. The tool is intended to use in feature vector calculation of region of interest (ROI) for real time video applications. These ROIs arrives serially. Also, is possible to evaluate the vector in design time, allowing system prototyping. The compiler optimizes the response time and the number of registers required to meet real time constraints.

  • 关键词:MISD; feature vector computation; compiler; video processing
国家哲学社会科学文献中心版权所有