RETREVIAL OF THE ORIGINAL IMAGE BY ARTIFICIAL NEURAL NETWORK

被引:0
作者
Sankaran, K. Sakthidasan [1 ]
Srinithya, G. [1 ]
Nagarajan, V. [1 ]
机构
[1] Adhiparasakthi Engn Coll, Elect & Commun Dept, Melmaruvathur, Tamil Nadu, India
来源
2014 INTERNATIONAL CONFERENCE ON COMMUNICATIONS AND SIGNAL PROCESSING (ICCSP) | 2014年
关键词
Artificial neural network; Levenberg marquardt algorithm; restoration; Noise level; Edge preservation; SURE;
D O I
暂无
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Restoration along with noise removal is a challenging task in image processing. The image quality is degraded because of the variety of noises. Such noises are removed during restoration in order to obtain the original image as possible. But the removal of noise itself is a major drawback since depicting the type of the noise and the amount of noise is the toughest task. There exist so many filtering and non-filtering techniques for removing the noise. But each of these techniques has several advantages and drawbacks. Filtering techniques involves various filters for restoration and the non filtering techniques are the various algorithms and the artificial neural network methodology used for the restoration. The proposed technique involves using the levenberg marquardt algorithm for restoration. The simulation results shows the usefulness of the proposed method.
引用
收藏
页数:5
相关论文
共 15 条
[11]  
Wilamowski B. M., 2002, MECHATRONICS HDB
[12]  
Wilamowski BM, 2003, IEEE IJCNN, P2546
[13]   Improved Computation for Levenberg-Marquardt Training [J].
Wilamowski, Bogdan M. ;
Yu, Hao .
IEEE TRANSACTIONS ON NEURAL NETWORKS, 2010, 21 (06) :930-937
[14]   Neural Network Architectures and Learning Algorithms [J].
Wilamowski, Bogdan M. .
IEEE INDUSTRIAL ELECTRONICS MAGAZINE, 2009, 3 (04) :56-63
[15]   Automatic Parameter Selection for Denoising Algorithms Using a No-Reference Measure of Image Content [J].
Zhu, Xiang ;
Milanfar, Peyman .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 2010, 19 (12) :3116-3132