Multi-target recognition with improved BP algorithm

XU Ting-fa; ZHANG Min; GU Hai-jun; SONG Jian-zhong; HAN Guang-liang

doi:null

您当前的位置：

首页 >

文章列表页 >

Multi-target recognition with improved BP algorithm

更新时间：2020-08-12

- Multi-target recognition with improved BP algorithm
- Optics and Precision Engineering Vol. 11, Issue 5, Pages: 513-515(2003)
- 作者机构：
  
  1. 中国科学院, 长春光学精密机械与物理研究所,吉林长春,130033
  2. 长春师范学院, 数理学院,吉林长春,130024
- 作者简介：
- 基金信息：
- DOI：
  CLC： TP183;TP391.41
- Received：06 March 2003，
  
  Revised：18 July 2003，
  
  Published Online：15 October 2003，
  
  Published：15 October 2003
- 稿件说明：
移动端阅览
许廷发, 张敏, 顾海军, 宋建中, 韩广良. 改进的BP算法在多目标识别中的应用[J]. 光学精密工程, 2003,(5): 513-515

XU Ting-fa, ZHANG Min, GU Hai-jun, SONG Jian-zhong, HAN Guang-liang. Multi-target recognition with improved BP algorithm[J]. Editorial Office of Optics and Precision Engineering, 2003,(5): 513-515
许廷发, 张敏, 顾海军, 宋建中, 韩广良. 改进的BP算法在多目标识别中的应用[J]. 光学精密工程, 2003,(5): 513-515 DOI：

XU Ting-fa, ZHANG Min, GU Hai-jun, SONG Jian-zhong, HAN Guang-liang. Multi-target recognition with improved BP algorithm[J]. Editorial Office of Optics and Precision Engineering, 2003,(5): 513-515 DOI：

摘要

为了实现对多目标的识别

提出了一种改进的BP算法。采用在步长函数中加入变动量因子

即当连续两次迭代其梯度方向相同时

步长加倍;当连续两次迭代其梯度方向相反时

步长减半。由于步长在迭代过程中自适应进行调整

使误差函数E在超曲面上的不同方向按照步长向极小点逼近

实现了对目标函数的优化。给出了在目标函数最优时的BP网络对三种飞机测试集的识别结果

其收敛速度比传统BP算法快4倍以上

表明该方法能够有效地用于多目标的识别。

Abstract

An improved BP algorithm has been proposed for realization of multi-target recognition. The principle followed is addition of an alternating momentum factor to the step function

i.e.

the step length is doubled when the gradient directions of two consecutive iterations are the same

and the step length is halved when the gradient directions of two consecutive iterations are opposite. Due to the adaptive adjustment of step length in the iteration process

error function E approaches the minimum point in different directions on the super curved surface in accordance with the step length to realize the optimization of goal function. The recognition rates achieved using the improved BP algorithm with optimized goal function are better than 98%

and the improved BP algorithm's convergence speed is four-times fast than the convergence speed of traditional BP algorithm.

关键词

Keywords

references

阎平凡,张长水.人工神经网络和模拟进化计算[M].清华大学出版社,2001.YAN P F, ZHANG CH SH. Artifical neural network and simulation evolve compute[M]. Tsinghua University Compress,2001.(in Chinese)

KEN T, NEIL D. The recognition and analysis of animate objects using neural networks and active contour models[J]. Neurocomputing,2002(43):145-172.

周建华,白钊.光度信号的频域模式识别方法[J].光学精密工程,2002(4):346-353.ZHOU J H, BAI ZH. Modal identification of luminosity signal in frequency domain[J]. Optics and Precision Engineering,2002,10(4):346-353.(in Chinese)

孙圣和.改进的非线性最小二乘法算法训练多层前馈神经网络[J].电子学报,1997,25(1):124-127.SUN SH H. Modified nonlinear least square algorithm for training multilayer feedforward neural networks[J]. Acta Electronic Sinica,1997,25(1):124-127.(in Chinese)

杨大力.多层前向神经网络中BP算法的误调分析及其改进的算法[J].电子学报,1995,23(1):117-120.YANG D L. The misadjustment analysis of BP algorithm and an improved algorithm[J]. Acta Electronic Sinica,1995,23(1):117-120.(in Chinese)

Views

440

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Full-stokes photodetector based on neural networks

Detection of conductive multi-particles based on circular convolutional neural network

A method for dense occlusion target recognition of service robots based on improved YOLOv7

Partial optimal transport-based domain adaptation for hyperspectral image classification

Saliency maps for point cloud recognition models

Related Author

WANG Shoutong

ZHANG Ran

CHU Jinkui

MENG Hailong

CAI Dehao

LUO Chen

ZHOU Yijun

JIA Lei

Related Institution

Liaoning Provincial Key Laboratory of Micro Nanosystems， Dalian University of Technology

Ningbo Research Institute， Dalian University of Technology

College of Mechanical Engineering， Southeast University

Wuxi Shangshi-finevision Technology Co.， Ltd

Chongqing Engineering Laboratory of Traffic Engineering Application Robot， Chongqing Jiaotong University

AI问答

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China Postal code：130033
Tel：0431-86176855 Email：gxjmgc@ciomp.ac.cn
Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰