吉林大学学报(信息科学版)

• 论文 • 上一篇    下一篇

基于非负矩阵分解的阴影检测方法

周鹏宇, 杨欣, 周大可, 刘加   

  1. 南京航空航天大学 自动化学院, 南京 210016
  • 出版日期:2013-11-26 发布日期:2014-01-06
  • 作者简介:周鹏宇(1989—), 男, 甘肃嘉峪关人, 南京航空航天大学硕士研究生, 主要从事数字图像处理、 人体行为识别研究, (Tel)86-25-84892305-5112(E-mail)xinyue042510225@163.com; 杨欣(1978—), 男, 江苏镇江人, 南京航空航天大学副教授, 硕士生导师, 主要从事智能控制、 模式识别和计算机视觉研究,(Tel)86-25-84892305-5112(E-mail)yangxin@nuaa.edu.cn。

Method of Shadow Detection Based on Non-Negative Matrix Factorization

ZHOU Peng-yu, YANG Xin, ZHOU Da-ke, LIU Jia   

  1. College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China
  • Online:2013-11-26 Published:2014-01-06

摘要:

针对以往的矩阵分解方法不能保证分解结果非负的问题, 根据非负矩阵分解(NMF: Nonnegative Matrix Factorization)结果非负的特点, 提出了基于NMF的阴影检测方法, 并以此为基础将进一步引入的分块非负矩阵分解(BNMF: Block Nonnegative Matrix Factorization)应用于阴影检测。通过NMF/BNMF提取训练样本中阴影的亮度特征, 再根据特征识别测试样本中的阴影区域。实验结果表明,  与基于奇异值分解方法相比, 该算法的阴影检测细节更清晰, 具有更好的效果。

关键词: 阴影检测, 非负矩阵分解, 分块非负矩阵分解

Abstract:

Previous matrix factorization algorithms can not guarantee the nonnegativity of results. Inspired by the nonnegative character of NMF (Non-negative Matrix Factorization), NMF is utilized to detect shadow areas. The BNMF (Block Nonnegative Matrix Factorization) is introduced for shadow detection. The brightness features of shadow points of training samples are extracted by NMF or BNMF method. The shadow areas of testing sample are recognized by features. The results demonstrate that the approach preserves edges clearer than the method based on singular value decomposition.

Key words: shadow detection, non-negative matrix factorization (NMF), block non-negative matrix factorization (BNMF)

中图分类号: 

  • TP391