Dissertation > Excellent graduate degree dissertation topics show

The Method of Image De-noising Based on Multi-dictionary and Sparse Representation

Author: WangZiZuo
Tutor: JinTaiSong
School: Xiamen University
Course: Computer Software and Theory
Keywords: Sparse Representation Image Denoising Multiple Dictionary
CLC: TP391.41
Type: Master's thesis
Year: 2014
Downloads: 6
Quote: 0
Read: Download Dissertation

Abstract


Image may be contaminated noises during each phase in image process applications, e.g.,acquisition, coding and transmission. It significantly decreases the image quality. Thus, image de-noising has always been a hot research topic.Sparse representation has been shown to be more effective than traditional de-noising methods due to adapting to physical environment, less sampling rate and data processing efforts, which significantly enhance the performance of de-noising. In view of this, scholars in related fields have attached great importance to the theory research based on completed atom signal sparse decomposition studies, which has a rapid development in recent years and become a mainstream signal representation. In the field of image de-noising applications, we have made some progress. Firstly, a brief overview is given in this paper on the traditional image de-noising such as classical wavelet de-noising.Then, we explore further the sparse correlation algorithm and techniques based on the initial DCT dictionary.A novel algorithm based on K-SVD and multi-dictionary is proposed. The main contents are as follows:(1)The overview on the image de-noising algorithms and the brief introduction of the over-complete sparse decomposition theory are denoted.(2)We deeply analyzes methods based on orthogonal matching pursuit, Orthogonal Matching, etc. Specifically the complexity and the detail of dictionary updating and signal reconstruction are focused. Furthermore, we describe the dictionary updating algorithm for optimization, named K-SVD.(3)A novel algorithm is proposed based on K-SVD and multi dictionary. This algorithm exploits the initial DCT over-complete dictionary constructed by running K-SVD algorithm on the noisy image and obtains an efficient representation of images, this thesis introduces the idea of multiple-dictionary. Experimental results demonstrate that Compared with the previous algorithms, our proposed algorithm performed better than the traditional denoising algorithm and the sparse representation based denoising algorithm on single DCT redundant dictionary. In addition, our proposed algorithm can preserves more information of original image.(4)The detail of discuss on Analysis K-SVD(Analysis K-means Singular Value Decomposition) algorithm is presented, the experimental results showed its significance.

Related Dissertations

  1. Research on Signal Recoveryalgorithms Based on Compressed Sensing,TN911.7
  2. The Fatigue Estimation and Real-Time Monitoring Based on Eeg,TN911.6
  3. Dictionary sparse representation based on adaptive super-resolution reconstruction of video coding technology,TN919.81
  4. Weak sparse underdetermined blind signal separation technology research,TN911.7
  5. Denoising problem solving variable weight fixed point algorithm,O177.91
  6. Training and weighted online dictionary sparse representation of difference,TP391.41
  7. Image Denoising Based on Statistical Model and Non-local Means Algorithm,TP391.41
  8. Infant Pain Facial ExpressioN Recogintion Based on Compressive Sensing,TP391.41
  9. Research on Technology of Image Segmentation Based on Level Set Method,TP391.41
  10. Speech signal based on sparse representation underdetermined blind separation,TN912.3
  11. Research on Image Denoising Based on Multiscale Geometric Transform,TP391.41
  12. The Structure of One Class of Signal Compression Algorithm,TP301.6
  13. Research on Feature Selection Algorithms Based on Pairwise Constraints and Sparse Representation,TP301.6
  14. Research of Gravid Ultrasonic Image Processing Based on Wavelet Transform,TP391.41
  15. Reconstruction Algorithms for Compressive Sensing and Their Applications to Digital Watermarking,TP309.7
  16. Wavelet-based image denoising method,TP391.41
  17. Study on Methods of PDEs about Image Processing,TP391.41
  18. Research of Nonsubsampled Contourlet Transform and Its Application on Image Processing,TN911.73
  19. Image Denoising Method Based on Wavelet Transform,TP391.41
  20. Wavelet - based remote sensing image super-resolution reconstruction,TP751
  21. Image Denoising Methods Based on PDE and Graph-cut,TP391.41

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net  Mobile