Dissertation > Excellent graduate degree dissertation topics show

Research on Medical Image Enhancement and Registration Similarity Measures

Author: ChenBeiJing
Tutor: DongGuangChang
School: Zhejiang University
Course: Applied Mathematics
Keywords: Medical Image Registration Order mutual information Second-order mutual information Local extrema Multi-resolution
CLC: TP391.41
Type: Master's thesis
Year: 2006
Downloads: 632
Quote: 5
Read: Download Dissertation

Abstract


Medical Imaging has become an important part of modern medicine, and medical image processing image processing, handling, to facilitate the clinical diagnosis. This article focuses on two types of medical image processing technology: based medical image contrast enhancement technology and current research focus of medical image registration. , Is a simple but effective contrast enhancement method based on gray-scale transformation of the medical image contrast enhancement technology. After the common gray transform Overview for some algorithms existing image contrast enhancement and contradictions between the edge detail to keep, we propose two new MR image contrast enhancement algorithm: based on threshold segmentation and B-spline interpolation of MR enhancement algorithms, first defined Otsu threshold algorithm selects multiple threshold to seek the best cubic B-spline interpolation transform the quality evaluation parameters based on the contrast and detail a new quality evaluation parameters, then, the ideal enhancement ; based segmentation and cumulative index transform the MR enhancement algorithms, the first use of multi-threshold segmentation image segmentation into different regions, then the statistics of the mean and variance of each region by the regional cumulative index nonlinear constructed The conversion function at all Regional enhancements. The experiments show that these two algorithms can be a better solution to contrast enhancement and edge detail to keep the two aspects of the conflict, and the speed as much as some of the traditional algorithms. Based on gray-scale image registration technology to achieve automatic registration is widely used for its high accuracy, no pretreatment. The similarity measure is to determine the alignment accuracy, robustness and real-time the most important factor, so in the introduction to the Registration and involved strong robustness of mutual information for first-order correlation analysis, and introduced several improvements mutual information measure, and focus on particular order mutual information, grayscale, relevant information and noise impact, the best order mutual information parameters, and then analyze the impact of the resolution of the first-order and second-order mutual information to improve the registration of two resolution strategies, using different similarity measure in different resolution level. The experiments show that, compared to a single measure of the algorithm, multi-measure algorithm combining both registration accuracy and speed can be more desirable effect.

Related Dissertations

  1. Research and Implementation of Non-rigid Medical Image Registration Method Based on Improved Optical Flow Model,TP391.41
  2. Algorithm and Implementation Based on Non-rigid Medical Image Registration,TP391.41
  3. Based on multi-resolution clustering algorithm for secure localization,TN929.5
  4. Medical image registration platform and combines new gray with geometric information Registration Measures,TP391.41
  5. Multi- resolution modeling key technology research,TP391.9
  6. Infrared and visible image registration and fusion research,TP391.41
  7. CT and ultrasound data registration studies and implementation,TP391.41
  8. Research and Implemente of Partial Discharge Location System,TM41
  9. Research on Acceleration Technique of 3D Terrain Display on Embedded System,TP391.41
  10. Construction of Multi-Band Orthogonal Multiwavelets with Multiplicity 2,O174.2
  11. The Study of Medical Image Enhancement Based on Wavelet,TP391.41
  12. Research on the Method of Enhancement of Fog Road Image,TP391.41
  13. Research on Feature-driven Image Registration and Image Mosaic of CT Images,TP391.41
  14. Based on multi- directional contrast image fusion research,TP751
  15. Some studies have valued wavelet,O174.2
  16. Enhancement Algorithm of Image Based on Retinex Theory in the Wavelet Domain,TP391.41
  17. Based on Wavelet Change’s Vehicle Identification System Research,TP391.41
  18. Some Studies on M-Band Wavelets and Wavelet Frames,O174.2
  19. The improved intelligent algorithm - based electric locomotive traction transformer fault diagnosis technology research,U269.6
  20. The Syudy on Alogrithms of Robust Image Watermarking Based on Transform Domain,TP309.7
  21. Several studies on the bidirectional wavelet,O174.2

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