Dissertation > Excellent graduate degree dissertation topics show

Research on Image Denoising Based on Multiscale Geometric Transform

Author: FuZhongKai
Tutor: WangXiangYang
School: Liaoning Normal University
Course: Applied Computer Technology
Keywords: Denoising Multiscale Geometric Threshold Support Vector Machine Plural direction pyramid transform (PDTDFB)
CLC: TP391.41
Type: Master's thesis
Year: 2010
Downloads: 75
Quote: 0
Read: Download Dissertation

Abstract


The digital image in the transfer and retrieval process, often contaminated by noise, and the subsequent processing to the image, such as segmentation, encoding, etc. of the image denoising first becomes very necessary. The wavelet transform is a new developed following the Fourier transform of a multi-resolution characteristics transform. Wavelet has a frequency focused, multi-resolution, low redundancy Porgy rich features, make it very suitable for image denoising. For images, however, the edge of the discontinuity is in accordance with the spatial distribution of the singularity of wavelet expansion progression wavelet expansion coefficients are not sparse, thus affecting the approximation error. Order to effectively overcome the deficiencies of wavelet transform denoising, people turned to a new type of singularity analysis tools - multi-scale geometric transformation. This paper studies the geometric transformation based on multi-scale image denoising main work is as follows: 1. Transformed into superior performance nonsubsampled contour based on a new image denoising method. Firstly the image nonsubsampled contour Transform, to obtain a different scale, the transform coefficients in the different directions; then combined with the noise distribution to determine the characteristics of the multi-scale threshold value, and so the threshold value for the high frequency coefficients denoising; last on noise processing after the transform coefficients to the inverse transform, to obtain the denoised image. The method not only has a strong ability to suppress noise, but also has better edge protection capabilities, while eliminating the phenomenon of the image near the edge of the pseudo-Gibbs (Gibbs). Due to the limitations of the threshold denoising method, we studied the image denoising method based on support vector machine. Firstly, the image wavelet transform to transform coefficients in different directions on different scales; then combined noise distribution in each direction on each scale to determine the characteristics of the spatial characteristics of the coefficients, and in turn the spatial characteristics of tectonic least squares support vector the training feature vectors of the machine; Finally, for all of the wavelet coefficients to be classified with the LS-SVM training of noise and non-noise signal, and so classification of high frequency coefficients denoising, and the denoising processing transform coefficients inverse transform denoising image. 3 due to the limitations of the wavelet, in recent years, multi-scale geometric transformation is widely applied in the various aspects of the image processing. This paper studies the plural direction translation invariant the space pyramid transform domain (PDTDFB) and direction Denoising scale model of adaptive Gaussian mixture.

Related Dissertations

  1. Research on Electromagnetic Acoustic Transducer Detection System Based on FPGA,TH878.2
  2. Performance Analysis of Anti-Multipath Capability in DSSS System and Study on the Compensation Methods,TN914.42
  3. Research on Autamatic Music Structrue Analysis,TN912.3
  4. Research on Methods of Medical Ultrasound Image Denoising,TP391.41
  5. Research on Transductive Support Vector Machine and Its Application in Image Retrieval,TP391.41
  6. Process Support Vector Machine and Its Application to Satellite Thermal Equilibrium Temperature Prediction,TP183
  7. Design of Testing Equipment for Electronic Products,TN06
  8. Research for Infrared Image Target Identification and Tracking Technology,TP391.41
  9. The Research on Intrusion Detection System Based on Machine Learning,TP393.08
  10. The Study of Human Face Detection Method Based on OpenCV,TP391.41
  11. Based on a sample study of the finite element mesh model optimization,TB114
  12. Based on Information Fusion roads and obstacle detection method,TP242
  13. Solder paste printing quality based on image processing techniques to detect,TP391.41
  14. Signal Processing Method Based on Singular Value Decomposition and Its Application to Mechanical Fault Diagnosis,TH165.3
  15. Research on Problems of Pumping-jack Diagnosis Based on Support Vector Machine,TP18
  16. Based on Information Fusion Fault Diagnosis of Analog Circuits,TN710
  17. Denoising problem solving variable weight fixed point algorithm,O177.91
  18. Customer value assessment based on comprehensive foreign trade enterprise channels of customer classification,F752;F224
  19. Based on SVM-RFE potential biomarker selection algorithm,TP311.13
  20. Rural road environment obstacle detection,TP391.41
  21. Machine learning based on sparse coding and image content recognition algorithm,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