Dissertation > Excellent graduate degree dissertation topics show

Learning-based image super-resolution technology and its application

Author: YangFang
Tutor: ShenJianBing
School: Beijing Institute of Technology
Course: Computer Science and Technology
Keywords: Super-resolution Energy Optimization Based on learning Gradient-based Markov networks
CLC: TP391.41
Type: Master's thesis
Year: 2011
Downloads: 94
Quote: 0
Read: Download Dissertation

Abstract


Super-resolution imaging technology's main purpose is to provide a one or more images from the same scene (information similar but different in detail) low resolution (Low-Resolution, LR) image reconstruction resolution (High-Resolution, HR) image, the image acquisition process which overcomes the limitations pathological conditions, and to obtain better image content, improve the accuracy of the scene recognition. This technology is widely used in remote sensing recognition, image compression, high-definition television, security monitoring, video, communications, medical diagnostics, resource exploration and many other areas, is the current field of image processing and computer vision in one of the most popular research direction, have a very important theoretical research value. This paper studies the learning-based image super-resolution algorithm and gradient-based image super-resolution algorithm, and image detail enhancement related knowledge, presented at the energy optimization framework, combined with edges and details of learning-based single image super- resolution algorithms. In this article first introduces the concept of super-resolution images, topic background, significance, and nearly a decade of domestic and international developments. Secondly describes the theoretical basis of the super-resolution technology and super-resolution algorithms on the airspace and a brief analysis of the advantages and disadvantages of various algorithms. While using Matlab tool to achieve a correlation algorithm to obtain the enlarged image. Then study the Markov network to learn the high-resolution images and the relationship between the low-resolution image, and through this relationship for instance-based super-resolution image super-resolution algorithms and the use of image gradient sharpening degree of prior knowledge to guide the gradient-based super-resolution image super-resolution, and then through the actual programming to achieve these two algorithms to obtain the corresponding experimental results. Through theoretical analysis of the experimental results that can be combined with edges and details two aspects to improve after a super-resolution image quality after finally propose a new single image super-resolution algorithm. This paper focuses on energy optimization framework based on learning of single image super-resolution algorithm, chapter 3.3.1 gives a super-resolution model, and define the energy equation and solving methods, chapter 3.3.2 describes the image itself training set to obtain, chapter 3.3.3 describes the search method, chapter 3.3.4 of the super-resolution algorithm is given in pseudocode flowchart. As can be seen from the experimental results, the proposed algorithm obtained high-resolution images at the edges and details are improved, with better visual effects. Finally, the content of this study are summarized, super-resolution face some problems to solve the problems and future research priorities and the direction of its future development were discussed.

Related Dissertations

  1. Research on the Super-resolution Technologies of Polarimetric SAR Images,TN957.52
  2. Superresolution of Hyperspectral Images Based on Spatial-Spectral Information Coordination,TN911.73
  3. Research on Regression-based Super-Resolution Image Reconstruction Technique,TP391.41
  4. Research on Image Super-resolution Reconstruction Based on Non-local Similarity,TP391.41
  5. Research on Color Image Super Resolution and Parallel Processing Technology,TP391.41
  6. Research on Learning-Based Super-Resolution and SR System Implementation,TP391.41
  7. Research on Far-field Subwavelength Focusing,O439
  8. Path Length Coded Coherence Combination for OCT Three-dimensional Superresolution,TP391.41
  9. Application Study of Compressed Sensing in Image Processing,TP391.41
  10. Image and Video Compression Techniques Based on Super-Resolution Reconstruction Algorithm,TN919.81
  11. Research for Super-resolution Reconstruction Technologies of Image Based on POCS Algorithm,TP391.41
  12. Research of Subwavelength Focusing Effect of Light by a Tapered Micro-nano Tube and Its Application,TN25
  13. Dictionary sparse representation based on adaptive super-resolution reconstruction of video coding technology,TN919.81
  14. Tetrolet sparse regularization learning with the sample image super-resolution algorithm,TP391.41
  15. Based on the sample image quality enhancement,TP391.41
  16. Stepped frequency SAR super-resolution imaging technology,TN957.52
  17. Face Image Super-Resolution Technology,TP391.41
  18. Research on Video Image Super Resolution Algorithm,TP391.41
  19. Based on regularization super-resolution image sequence reconstruction,TP391.41
  20. Research on Algorithm of Image Super-resolution Reconstruction,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