Dissertation > Excellent graduate degree dissertation topics show

A Method for Image Quality Assessment Based on Visual Attention and Structure Distortion

Author: XiangBin
Tutor: LiuWenYu
School: Huazhong University of Science and Technology
Course: Communication and Information System
Keywords: Quality assessment Visual attention Human visual system Saliency map Structural distortion
CLC: TP391.41
Type: Master's thesis
Year: 2011
Downloads: 30
Quote: 0
Read: Download Dissertation

Abstract


The digital image quality evaluation in the field of computer vision and machine learning is a fundamental and important research topics in the areas of image quality control, image processing algorithm standard visual system and algorithm parameters optimization has broad application prospects. Digital image quality evaluation is also a problem in the computer vision field, this is because in the process of image acquisition and image processing such as compression and transmission will produce a variety of different types of distortion, and makes it difficult to have a unified evaluation method to solve the various types of distortion problems . The same time, the quality assessment is a visual perception problems, visual perception is a very subjective and physiological mechanisms, it can not make a specific description. So give objective and accurate assessment of the image quality is a difficult problem. Performance? Explicit introduction of the traditional quality evaluation method of visual attention mechanism so as to enhance the original method is the goal of this study. The traditional algorithm error statistics and HVS model-based algorithm. Focus only on the error statistics-based approach to measure the difference between the distorted image with the reference image, without considering the human visual perception of the influence of these differences, and therefore have greater access and between the subjective evaluation value. HVS model-based algorithm by the characteristics of the a comprehensive simulation HVS system, better able to ensure the consistency of the subjective and objective evaluation, however, the complexity of the algorithm, the large amount of computation, and is subject to the current level of awareness of HVS mechanisms still need further research. In this paper, the combination of these two ideas, visual attention mechanism algorithm based on error statistics, which play the advantages of both, to improve the performance of the algorithm. The paper first introduces the basic concepts and theoretical basis of quality assessment and framework, and propose a new quality evaluation system for the introduction of visual attention mechanisms, and theoretical analysis of visual salience reasonable to enhance the performance of the algorithm; respectively, after the introduction of this article the two basic algorithms: based on the context of the visual significantly with Spectrum extraction algorithms and based on structural similarity of image quality evaluation algorithm, and carried out the algorithm analysis and experimental validation; Then, for visually significant Spectrum and error distortion map the characteristics and design of the The combination of three methods, so that the two maps can be combined, after experimental verification, the two spectra are multiplied and normalized effective, can greatly enhance the original algorithm performance. ?

Related Dissertations

  1. Research on Sparse Image Representation and Coding Model,TP391.41
  2. Effect of Soil Erosion on Soil Quality in the Cultivated Slope Land on the Loess Plateau, China,S157.1
  3. Research on the Early-warning Technical System for the Water Environment of the River Network,X853
  4. Investigation and Assessment of Safety on Vegetable Soil Fertiltiy and Environment Quality in Suburbs of Nanjing,S631.1
  5. Xinjiang Chaiwopu Lake 2005 - 2009 water quality status and trends,X832
  6. Research on NO-reference Image Quality Assessment Based on HVS,TP391.41
  7. Research of Scalable Video Coding Based on H.264/AVC Extension,TN919.81
  8. Research of Visual Attention Model for Complex Object,TP391.41
  9. Distortion effects on image quality evaluation and classification,TP391.41
  10. Characteristic style of clothing based on cognitive research and implementation of self- classification,TS941.11
  11. Xi'an Qujiang Pool Water Quality Assessment and Water Quality Simulation,X824
  12. Xi'an moat Analysis and Control of Water Pollution explore ways,X522
  13. Environmental audit quality indicators and methods of evaluation,F239.6
  14. Small mobile online water quality monitoring system,TP274
  15. EPA and reliability of the wireless link quality assessment mechanism of maintenance,TP273
  16. Image Quality Evaluation Based on HVS research,TP391.41
  17. Electronic passport quality assessment system,TP391.41
  18. Based on the structural characteristics of human visual similarity image quality evaluation,TP391.41
  19. Saliency -based image retrieval keying Research and Implementation,TP391.41
  20. Low-quality document image binarization research,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