基于HPI的神经网络图像匹配多处理机系统

石争浩; 冯亚宁; 张遂南; 黄士坦

基于HPI的神经网络图像匹配多处理机系统

1.
西安理工大学计算机科学与工程学院西安 710048
2. 西安微电子技术研究所, 西安 710054

基金项目: 航天"十五"预研基金资助项目(413160203); 西安理工大学中青年创新基金资助项目(116-210302)

详细信息

作者简介:
石争浩 (1968-),男,陕西富平人,博士生, fyn-szh@263.net.

中图分类号: TP 391
计量
- 文章访问数: 2876
- HTML全文浏览量: 20
- PDF下载量: 872
- 被引次数: 0
出版历程
- 收稿日期: 2004-06-25
- 网络出版日期: 2004-11-30

HPI based multi-processor system for neural network image matching

1.
School of Computer Science, Xi'an University of Technology, Xi 'an 710048, China
2. Xi'an Institute of MicroElectronics, Xi'an 710054, China

摘要

摘要: 针对嵌入式图像匹配计算特点,采用TMS320C6X系列处理器作为并行神经处理单元,设计了一种基于TMS320C6X系列处理器HPI(Host-Port-Interface)互连的神经网络图像匹配多处理机系统,在这种并行计算系统中,包括一个主控计算单元和三个并行神经计算单元,主控计算单元通过HPI接口与各个神经匹配处理单元直接连接,通过HPI接口,主控计算单元可以直接访问各个神经元的片上和片外存储器,实现实时图像数据的直接转发和神经元中间运算结果的读取.理论分析表明,该设计可有效优化神经计算结构,提高图像匹配的实时性.
- 图像匹配 /
- 神经网络 /
- 多处理机 /
- HPI(Host-Port-Int erface)
Abstract: Aim at the characteristics of embedded image matching, TMS320C6X family processo rs were used for parallel neural computing units, and a multi-processo r system structure based on TMS320C6X HPI(host parasite interface) interlinkin g for neural network image matching was presented. In this parallel computing sy stem, one host computer and three DSP neural units are contained, the master com puter is directly linked with each neural computing unit by HPI, memory of each neural unit can be directly accessed by the master computer through their HPI, t he real time image datum and the inter-results datum of each neural unit can be immediately directed transmit to each other between neural unit and the master c omputer. As a result, the real time performance of image matching is effectively improved according to theoretical analysis.
- image matching /
- artificial neural network /
- multi-processor /
- HPI(host port inte rface)

HTML全文

参考文献(1)

[1] 李俊山,沈绪榜.图像匹配技术研究[J]. 微电子学与计算机,2000, 17(2):10~14 Li Junshan,Shen Xubang. Study on image matching[J]. Microelectronics and Computer, 2000, 17(2):10~14(in Chinese) [2] 石争浩,黄士坦,冯亚宁,等.人工神经网络图像匹配技术[J].微电子学与计算机,2003,20(增):18~21 Shi Zhenghao, Huang Shitan, Feng Yaning,et al. Artificial neural network image matching[J]. Microelectronics and Computer, 2003, 20(Supl):18~21 (in Chinese) [3] Mauduit N. Lneuro1.O:A piece of hardware LEGO for building neural network systems[J]. IEEE Trans on Neural Networks,1992,3(3):414~422 [4]Yang J. 2-D optical fully bipolar neural network using colour encoding[J]. In IEE Proc Optoelectron, 1995,142(2):76~79 [5] 陈虎,戴葵,胡守仁.环行阵列神经网络计算机系统[J].计算机研究与发展,1999,36(2):144~149 Chen Hu, Dai Kui, Hu Shouren. A cycle array neurocomputer[J]. Journal of Computer Research and Development, 1999,36(2):144~149(in Chinese) [6] 罗莉,何鸿君,胡守仁.数字神经元芯片的设计与应用[J].计算机研究与发展,1998,35(9):798~802 Luo Li, He Hongjun, Hu Shouren.Design and application of a digital neural network processor slice[J]. Journal of Computer Research and Development,1998, 35 (9):798~802(in Chinese) [7] 章锦文.人工神经网络信号处理机的设计与实现 .西安:西北工业大学,1994 Zhang Jinwen. Designing and implementation of an artificial neural network signal processor system . Xi'an:Northwestern Polytechnic University,1994(in Chinese) [8] 张遂南.嵌入式实时图像匹配计算机硬件技术研究 . 西安:西安微电子技术研究所, 2002 Zhang Suinan. Research on hardware technology for embedded real time image matching computer . Xi'an:Xi'an Institute of Micro-Electronics, 2002(in Chinese) [9] 胡超,郑南宁,高岩,等.MIMD体系结构上神经网络的实现技术 . 见:中国神经网络学会编.中国神经网络1993年学术大会论文集(上册) ．西安:西安电子科技大学出版社,1993.540~545 Hu Chao, Zheng Nanning, Gao Yan,et al. Neural network implement technology on MIME structure . In:China Neural Networks Council, ed. Xi'an:1993 China Conference on Neural Network . Xi'an:Xidian University Publishing House,1993. 540~545(in Chinese)

施引文献

资源附件(0)

访问统计