Title: Wavelet Based Data Hiding of DEM in the Context of Realtime 3D Visualization
1Wavelet Based Data Hiding of DEM in the Context
of Real-time 3D Visualization
- Presented By
- Khizar Hayat
2The Team
- Student
- Khizar Hayat
- Supervisor
- William Puech (LIRMM/ROB/ICAR)
- Co-supervisors
- Marc Chaumont (LIRMM/ROB/ICAR)
- Gilles Gesquiere (LSIS - University of Marseille)
- Concerned PhD student
- Philippe Amat
3The Problem
- How to achieve optimal real-time 3D visualization
in a client server environment in a scalable and
synchronized way compatible with the clients
computing resources and requirements?
100 Data
A fraction of Data
4The Proposed Solution
- Our Approach
- Discrete Wavelet Transform (DWT) of JPEG2000
- Data Hiding LSB based embedding
- Advantages
- Compression to reduce the amount of data
- Scalability - different levels of detail for
various - Platforms
- Users
- contexts
- Synchronization reduce the number of files
5Essential Concepts
- JPEG2000
- Data Hiding/ Steganography
- 3D Visualization
6JPEG2000
7DWT in JPEG2000
Let Sj 1D input signal (pixel row or pixel
column) Lk lowpass subband signal Hk
highpass subband signal Where 1 j n and 1
k n/2 Then
- Lossless (Daubechies 5/3)
Lossy (Daubechies 9/7)
Where a - 1.586134 b - 0.052980 c
0.882911 d 0.443506 With ß and ß, the
scaling parameters ß 0.812893 ß 1/ ß
S2n1 the first stage lifting outcomes S2n
the second stage lifting outcomes
8LSB Based Data Hiding
- Embed the message bits directly into the
least-significant bit plane of the cover image in
a deterministic sequence, e.g. PRNG - Advantages
- Easy to understand and comprehend ?popular
- High perceptual transparency
- Low degradation in the image quality
- High Hiding Capacity
- Disadvantages
- Low robustness to malicious attacks
- Vulnerable to accidental or environmental noise
93D Visualization
- Three files are necessary
- Texture
- Altitude DEM ? Fig. a
- (50m by IGN France)
- Geo-referential coordinates (longitude /
latitude) - Two steps for visualization
- Triangulation (Fig. b)
- Aerial Photograph draped onto the triangles for
3D visualization (Fig. c)
(a)
(b)
(c)
10The Proposed Method
11The Proposed Method
- RGB to YCrCb
- Texture.ppm
- Wavelet Transformation
- Lossy DWT of the YCrCb of Texture
- Lossless DWT of altitude (DEM)
- Insertion
- Trans DEM Trans Y plane of texture
- Correspondence
- Same level
- Factor of insertion (E)
- E m²/ N² (coefficients/ pixel)
- Block size 1/E ? 2 bytes per 32x32 block in our
case - LSB based Insertion with running a PRNG for pixel
allocation - Final Interleaved Image
12Results
Altitude Image (64x64 coefficients) 1 coefficient
16 bits
Texture Image (2048x2048 pixels) 1 Pixel 24 bits
Texture Image (A part magnified)
13Level 1 Transformation
Texture (2048x2048 pixels)
Altitude (64x64 coefficients)
14Level 3 Transformation
Texture (2048x2048 pixels)
Altitude (64x64 coefficients)
15Observations
Where
16Reconstruction from the Image of Approximation at
Level 3
Altitude Image (64x64 coefficients) 1 coefficient
16 bits
Texture Image (2048x2048 pixels) 1 Pixel 24 bits
Texture Image (A part magnified)
ATTENTION Only 1.6 of the initial data utilized
17Results
Altitude Image (64x64 coefficients) 1 coefficient
16 bits
Texture Image (2048x2048 pixels) 1 Pixel 24 bits
Texture Image (A part magnified)
183D visualization of the Altitude from the Image
of Approximaton
Level 1
Level 0 - all the information
Level 2
Level 3
193D navigation of the Constructed Images
All the data
Level 3 lowest suband data
ATTENTION Only 1.6 of the initial data utilized
20Conclusion
- Encouraging Results in the following contexts
- Compression
- Scalability and portability
- Synchronization
- Future Perspective
- Other aspects of JPEG2000
- Standard Software, e.g. OpenJPEG
- Geometric Wavelets
- Non uniform grid to decrease the number of
triangles - Exploration of Chrominance planes
21Research Papers
- Papers expected to be published
- K. Hayat, M. Chaumont, G. Gesquiere and W. Puech.
Visualisation 3D Temps-Réel à Distance de MNT par
Insertion de Données Cachées Basée Ondelettes. In
CORESA06, Caen, France, Nov. 2006. (Submitted in
May 2006). - K. Hayat, W. Puech, G. Gesquiere and M. Chaumont.
Wavelet Based Data Hiding of DEM In the Context
of Real-time 3D Visualization. In IST and SPIE
07 - Visualization and Data Analysis 2007
(EI108), San Jose, California, USA (to be
submitted before 17 July 2006).
22References
- A. Martin, G. Gesquiere, W. Puech and S. Thon.
Real Time 3D Visualisation of DEM Combined with a
Robust DCT Based Data-Hiding Method. In
Electronic Imaging, Visualization and Data
Analysis, SPIE, IST, San Jose, CA, USA, volume
6060, pages 60600G160600G8, Jan 2006. - W. Sweldens. The Lifting Scheme a New Philosophy
in Biorthogonal Wavelet Constructions. In
Electronic Imaging, Wavelet Applications in
Signal and Image Processing, SPIE, IST, San
Diego, CA, USA, volume 2569, pages 6879, Sep.
1995. - I. Daubechies and W. Sweldens. Factoring Wavelet
Transforms into Lifting Steps. Fourier Anal.
Appl., 4(3), 1998. - ISO/IEC 14492-1 Lossy/lossless coding of
bi-level images, 2000. - R. Raffin S. Thon and G. Gesquiere. Visualisation
3D de Feux de Forêts sur des Modèles Numériques
de Terrain de lIGN. In Rencontre GEO-RISQUE 2006
La cartographie des risques naturels,
Montpellier (France), Feb. 2006. - P. Meerwald and A. Uhl. A Survey of
Wavelet-Domain Watermarking Algorithms (2001).
Electronic Imaging, Security and Watermarking of
Multimedia Contents, SPIE, IST, San Jose, CA,
USA, Jan 2001. - A. Piva, F. Bartolini and R. Caldelli. Self
Recovery Authentication of Images in the DWT
Domain. In International Journal of Image and
Graphics, pages 149-165, Vol. 5, No. 1(2005). - J. Li. Image Compression - The Mechanics of the
JPEG2000. Microsoft Research, Signal Processing
2002. (Available from http//research.microsoft.c
om/jinl/paper_2002/msri_jpeg.htm). - M.D. Adams. The JPEG-2000 Still Image Compression
Standard (Last Revised 2005-12-03). ISO/IEC JTC
1/SC 29/WG 1 (N 2412) (Available from
www.ece.uvic.ca/mdadams)
23Thank You