Image Adaptive Watermarking Project Proposal - PowerPoint by epmd

VIEWS: 62 PAGES: 10

									Image Adaptive
Watermarking
Project Proposal
Ryan Johnson
EE6886 Multimedia Security Systems
Prof. Ching-Yung Lin
April 5, 2006
Outline

    Introduction of proposed project
    Project purpose and objectives
    Techniques and references
    Project plan
      Research
      Experimentation
      Result evaluation
  Preliminary schedule
Project Proposal

  Image Adaptive (IA)
   Watermarking
    Attempt to improve
     the conflict between
     the three image
     watermarking metrics
    Attempt to overcome
     problems with uniform
     images
       Utilizes human visual
        models to maximize
        watermark length and
        power.
Project
Purpose/Objectives
  Understand Human Vision Model
    Human visual system
    Just Noticeable Difference (JND)
  Compare IA watermarking schemes
    DCT vs. Wavelet Transform
    Spread Spectrum Technique (Cox et al)
  Compare Human Vision Models within
   watermarking schemes
Techniques
  Discrete Wavelet
   Transform
  Discrete Cosine
   Transform
  Human Vision Models
     Obtaining JND thresholds
     Applying human vision
      models
  Watermark Extraction
References

  Watson (NASA), Rubin (NYU)
    Human visual studies
  Podilchuk/Zeng (Purdue)
    IA watermarking based on DCT and wavelet
     transform
  Cao et al (Mississippi State Univ.)
    Redundant Wavelet Transform
Project Plan

    Research Tasks (Week 1)
     1. Human Vision Models
          How human vision works
              How the human vision models were created
          Proposed human vision models
          Can I model a human vision system?
     2. Research various IA watermarking
        schemes
Project Plan

     Experimentation/Implementation
     1. Implement IA watermarking schemes (Week 2 – 3)
           Spread Spectrum technique (Cox et al)
           Different IA schemes
           Different human vision models
     2. Evaluate results (Week 3 - 4)
           JND threshold testing/validation
           Test procedure - determine variables and minimize
     3. Report results (Week 5)
           Write report
Evaluating Results
  JND Thresholds
     Compare colleague test results to proposed human vision models

  Watermarking metrics
     Transparency
         Human comparison test
         PSF
     Robustness - Attacks
           JPEG Compression
           Rescaling
           Cropping
           Any combination
     Capacity
         Maximum intensity watermark
         Maximum length watermark
     Overall improvement over Spread Spectrum
         Compare results of images with large, smooth areas for each type of
          watermarking scheme
Questions?

								
To top