Title: Image inpainting with sparse prior
1Image inpainting with sparse prior
- Spencer Brady
- ECE 544 Project
- 11/30/2007
2Outline
- Introduction
- Definition of Inpainting
- Problem Statement
- Simplifying Assumptions
- Results of Different Approaches
- DCT
- Separable 9-7 Wavelets
- Contourlets
- Morphological Component Analysis
- PSNR Comparisons
3Introduction
- Inpainting estimates the missing pixels based on
some prior image model
Inpainting
Original Image
Reconstructed Image
4General Problem Statement
We model the true image (with no missing pixels)
as having a sparse representation in T where
5General Problem Statement
We wish to solve the following optimization
problem
From which we recover the inpainted image
6Simplifying Assumptions
- Replace l0 norm with l1 norm
- Replace coefficients with inverse transform
, - then do the minimization over instead of
7Simplified Problem Statement
Now we minimize over the set of images instead
of the set of coefficients
The inpainted image is nothing but
8How do we solve this?
- Recall
- Has a decoupled solution given by soft
thresholding
ai
bi
d
-d
9General Algorithm
- Iterative Soft Thresholding
- Initialize Parameters (threshold d, iterations N,
xhat) - Loop N times
- (a) Decompose/Threshold/Reconstruct
- - Compute transform of M(x)(1-M)xhat
- - Soft threshold the coefficients with d
- - Reconstruct xhat from thresholded coefficients
- 3) Set output to of M(x)(1-M)xhat
10DCT Approach
- Choose
- T the 2-D Discrete Cosine Transform
11DCT Results
N40 d6
PSNR 40.64
12DCT Results
N40 d10
PSNR 31.11
13DCT Results
N40 d35
PSNR 34.04
14DCT Results
N50 d60
PSNR 30.00
15DCT Results
N80 d250
PSNR 22.36
16DCT Results
N60 d28
PSNR 32.98
17DCT Results
N60 d25
PSNR 35.21
18DCT Results
N40 d15
PSNR 42.18
19Separable Wavelet Approach
- Choose
- T separable 2-D Wavelet Transform
- Cohen-Daubechies-Feauveau biorthogonal 9-7
wavelet with 7 levels of decomposition
20Separable Wavelet Results
N60 d15
PSNR 36.61
21Separable Wavelet Results
N80 d15
PSNR 27.76
22Separable Wavelet Results
N70 d15
PSNR 34.12
23Separable Wavelet Results
N70 d20
PSNR 29.93
24Separable Wavelet Results
N150 d100
PSNR 21.69
25Separable Wavelet Results
N80 d30
PSNR 31.33
26Separable Wavelet Results
N90 d15
PSNR 33.81
27Separable Wavelet Results
N90 d10
PSNR 41.91
28Contourlet Approach
- Choose
- T the Contourlet Transform
- 5 4 4 3 3 levels/directions
- pkva filters
29Contourlets Results
N40 d5
PSNR 39.71
30Contourlets Results
N50 d10
PSNR 27.62
31Contourlets Results
N90 d10
PSNR 36.13
32Contourlets Results
N90 d10
PSNR 32.38
33Contourlets Results
N70 d40
PSNR 15.79
34Contourlets Results
N70 d20
PSNR 33.13
35Contourlets Results
N70 d8
PSNR 31.22
36Contourlets Results
N40 d10
PSNR 42.10
37Morphological Component Analysis Approach
Cartoon component
Texture component
38Sparse Image Model
Cartoon component
Texture component
Images obtained from 1
39Problem Statement
We wish to solve the following optimization
problem
From which we obtain the inpainted image
40Simplified Problem Statement
Now we minimize over a set of images rather than
a set of coefficients (reducing the
dimensionality of the problem)
From which the inpainted image simply becomes
41How do we solve this?
- Recall
- Has the decoupled solution of soft thresholding
ai
bi
d
-d
42Proposed Algorithm
- Iterative Soft Thresholding
- Initialize Parameters (threshold d, iterations N)
- Loop N times
- (a) Update xn
- - Compute contourlet transform of M(x-xt)(1-M)
xn - - Soft threshold the coefficients with d
- - Reconstruct xn from thresholded coefficients
- (b) Update xt
- - Compute local DCT transform of M(x-xn)(1-M)
xt - - Soft threshold the coefficients with d
- - Reconstruct xt from thresholded coefficients
- 3) Set output to xt xn
43MCA Results
N40 d5
PSNR 41.24
44MCA Results
N40 d5
PSNR 30.44
45MCA Results
N60 d30
PSNR 32.52
46MCA Results
N60 d60
PSNR 23.07
47MCA Results
N40 d5
PSNR 14.14
48MCA Results
N40 d5
PSNR 26.12
49MCA Results
N40 d5
PSNR 29.12
50MCA Results
N40 d5
PSNR 45.93
51PSNR Comparisons
52Comment
53References
- 1 Elad, Starck, Querre, Donoho, Simultaneous
cartoon and texture image inpainting using
morphological component analysis, Applied and
Computational Harmonic Analysis, 19 (2005)
340-358 - 2 Starck, Elad, Donoho, Image decomposition
via the combination of sparse representations and
a variational approach, IEEE Transactions on
Image Processing, Oct 2005 - 3 Rudin, Osher, Fatemi, Nonlinear total
variation based noise removal algorithms,
Physica D 60 (1992) 259-268 - 4 Sardy, Bruce, Tseng, Block coordinate
relaxation methods for nonparametric signal
denoising with wavelet dictionaries, Journal of
Computational and Graphical Statistics (2000)
361-379 - 5 Donoho, Ideal spatial adaptation by wavelet
shrinkage, Biometrica (1994) 81, 3, pp.425-55 - 6 Do, Vetterli, The contourlet transform an
efficient directional multiresolution image
representation, IEEE Transactions on Image
Processing, vol. 14, no. 12, pp. 2091-2106, Dec.
2005
54Questions