| 基于独立分量分析的图像特征提取及去噪 |
| |
| 引用本文： | 范羚,吴小培,龙飞,张道信,郭晓静. 基于独立分量分析的图像特征提取及去噪[J]. 计算机工程与应用, 2003, 39(9): 107-109,126 |
| |
| 作者姓名： | 范羚 吴小培 龙飞 张道信 郭晓静 |
| |
| 作者单位： | 1. 安徽大学计算智能与信号处理教育部重点实验室,合肥,230039
2. 安徽大学计算智能与信号处理教育部重点实验室,合肥,230039;中国科技大学电子科学与技术系,合肥,230026 |
| |
| 基金项目： | 安徽省自然科学基金资助项目(编号:0043214) |
| |
| 摘 要： | 探讨了一种新的多元统计分析方法———独立分量分析在图像特征提取及噪声去除中的应用。利用基于信息传输最大原则的infomax算法对自然图像的基向量进行迭代学习。实验表明,提取的基向量在空间频率上具有方向性和局部性,描述了输入自然景物图像的边缘特征。结合软门限算子,实现了图像中高斯噪声的去除。
|
| 关 键 词： | 独立分量分析 特征提取 软门限算子 稀疏编码 |
| 文章编号： | 1002-8331-(2003)08-0107-03 |
Image Feature Extraction and Image Denoising Based on Independent Component Analysis |
| |
| Fan Ling Wu Xiaopei , Long Fei Zhang Daoxin Guo Xiaojing. Image Feature Extraction and Image Denoising Based on Independent Component Analysis[J]. Computer Engineering and Applications, 2003, 39(9): 107-109,126 |
| |
| Authors: | Fan Ling Wu Xiaopei Long Fei Zhang Daoxin Guo Xiaojing |
| |
| Affiliation: | Fan Ling 1 Wu Xiaopei 1,2 Long Fei 1 Zhang Daoxin 1 Guo Xiaojing 11 |
| |
| Abstract: | A new multidimensional signal processing method—Independent Component Analysis(ICA)is applied to im-age feature extraction and image denoising.Using the infomax(information maximization)algorithm of ICA implemented by a forward neural network,it gets the basis vectors of the natural images.The experiments show that the basis vectors are localized and oriented,and sensitive to lines and edges of varying thickness of the images.Then the Gaussian noise is reduced by applying a soft -thresholding operator on the separated independent components.The performance of this method is appealing. |
| |
| Keywords: | Independent Component Analysis(ICA) Feature Extraction Soft-thresholding operator Sparse coding |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
