计算机应用 ›› 2013, Vol. 33 ›› Issue (03): 656-659.DOI: 10.3724/SP.J.1087.2013.00656

• 多媒体处理技术 • 上一篇    下一篇

基于改进的快速稀疏编码的图像特征提取

尚丽1,2*,苏品刚1,周燕1,3   

  1. 1.苏州市职业大学 电子信息工程学院, 江苏 苏州 215104;
    2.中国科学技术大学 自动化系, 合肥 230026;
    3.苏州大学 电子信息学院, 江苏 苏州 215006
  • 收稿日期:2012-09-04 修回日期:2012-10-27 出版日期:2013-03-01 发布日期:2013-03-01
  • 通讯作者: 尚丽
  • 作者简介:尚丽(1972-),女,安徽砀山人,副教授,博士,主要研究方向:人工智能、模式识别、数字图像处理; 苏品刚(1971-),男,江苏苏州人,副教授,主要研究方向:毫米波焦平面成像、测控; 周燕(1980-),女,江苏苏州人,讲师,博士研究生,主要研究方向:语音信号处理、模式识别。
  • 基金资助:

    国家自然科学基金资助项目(60970058); 江苏省自然科学基金资助项目(BK2009131); 江苏省“青蓝工程”项目; 2010苏州市职业大学创新团队项目(3100125)。

Image feature extraction based on modified fast sparse coding algorithm

SHANG Li1,2*, SU Pin'gang1, ZHOU Yan1,3   

  1. 1.School of Electronic Information Engineering, Suzhou Vocational University, Suzhou Jiangsu 215104, China;
    2.Department of Automation, University of Science and Technology of China, Hefei Anhui 230026, China;
    3.School of Electronic and Information, Soochow University, Suzhou Jiangsu 215006, China
  • Received:2012-09-04 Revised:2012-10-27 Online:2013-03-01 Published:2013-03-01
  • Contact: SHANG Li
  • Supported by:

    The Study of Non-negative Sparse Coding and Its Applications in Image Processing of Millimeter-wave Focal Plane Imaging;The grants of Natural Science Foundation of Jiangsu Province of China, No.BK2009131

摘要: 考虑图像特征系数的最大化稀疏分布和特征基的正交性,在快速稀疏编码(FSC)模型的基础上,提出一种改进的FSC模型。该模型利用迭代法解决了基于L1范数的归一化最小二乘法和基于L2范数的约束最小二乘法的凸优化问题,能够实现完备基和过完备基的学习,有效提取出图像的最佳特征,且比标准稀疏编码(BSC)模型的收敛速度快。分别利用自然场景图像和掌纹图像作为训练数据进行特征提取测试,并进一步利用提取的特征基进行图像重构实验,同时与BSC模型的图像重构结果进行对比,实验结果证实了所提出的改进FSC模型能够快速、有效地实现图像的特征提取。

关键词: 快速稀疏编码, 最小二乘法, L1范数, L2范数, 特征提取, 图像重

Abstract: On the basis of the Fast Sparse Coding (FSC) model, considering the maximum sparse distribution of feature coefficients and the orthogonality of feature bases of an image, a Modified FSC (MFSC) model was proposed in this paper. This FSC algorithm was based on iteratively solving two convex optimization problems: L1-norm based regularized least square problem and L2-norm based constrained least square problem, and it can realize the learning of complete bases and overcomplete bases, as well as efficiently extract the features of images. Moreover, the convergence speed of FSC is quicker than that of Basic Sparse Coding (BSC). The images of natural scene and palmprint were used to test the property of FSC algorithm proposed by the authors in feature extraction, and then the extracted features were utilized to image reconstruction. Compared with reconstructed images obtained by BSC, the experimental results verify the validity of the modified FSC in quickly extracting image features.

Key words: Fast Sparse Coding (FSC), least square, L1-norm, L2-norm, feature extraction, image reconstruction

中图分类号: