Methods and apparatus for image restoration

• US 8,285,070 B2
• Filed: 04/07/2009
• Issued: 10/09/2012
• Est. Priority Date: 04/07/2008
• Status: Expired due to Fees

• Current Assignee: Trustees of Tufts College
• Original Assignee: Trustees of Tufts College
• Inventors: Panetta, Karen A., Agaian, Sos, Greenblatt, Aaron
• Primary Examiner(s): Rahmjoo, Mike
• Application Number:
  US12/419,632
• Publication Number:
  US 20090274386A1
• Time in Patent Office: 1,281 Days
• Field of Search: 382/266, 382/275, 382/164, 382/261, 382/263, 382/109, 382/113, 382/118, 382/254
• US Class Current: 382/266
• CPC Class Codes: G06T 2207/20192 Edge enhancement; Edge preservation
• CPC Class Codes: G06T 5/77 Retouching; Inpainting; Scratch removal

Abstract

Methods and apparatus for restoration of a digital image. In one embodiment, a method for repairing a defect in a digital image to provide a restored image comprises determining a plurality of pixel locations to form a neighborhood relating to the defect and whether or not the neighborhood has a well-defined, dark border along its edge. Should the neighborhood not have dark border, one embodiment of the method entails processing the neighborhood to bring the neighborhood approximately to uniform darkness, processing the neighborhood to match surroundings in the digital image, copying an edge of a neighborhood in the digital image into the processed neighborhood, processing pixels of the edge to repair the copied edge pixels, and outputting the restored image for display to a user. Should the neighborhood have a dark, well-defined border, one embodiment of the method entails processing the neighborhood as to locally enhance the neighborhood and match its surroundings in the digital image; processing the neighborhood'"'"'s edge such that the edge also matches its surroundings in both the defect and the digital image; processing the neighborhood to invert its pixel values and then perform the last two steps once again; processing the neighborhood to increase its contrast and then perform the last three steps once again; processing the neighborhood to bring the neighborhood to a more uniform darkness; processing the uniform darkness neighborhood to match surroundings in the digital image; and outputting the restored image for display to a user.

Claims

1. A method for repairing a defect in a digital image to provide a restored image, comprising:

determining, using a computer having a processor, a plurality of pixel locations relating to the defect and whether a blotch has an edge darker than a selected threshold;

where it is determined that the edge is not darker than the selected threshold,

(a) processing the blotch to bring the blotch to a more uniform darkness;

(b) processing the uniform darkness blotch to match surroundings in the digital image;

(c) copying a border of the blotch in the digital image into the processed blotch;

(d) processing pixels of the border to repair the copied border pixels; and

(e) outputting the restored image for display to a user;

where it is determined that the edge is darker than the selected threshold,

(f) processing the blotch to locally enhance the blotch and match surroundings in the digital image;

(g) processing the blotch edge such that the edge also matches surroundings in the defect and the digital image;

(h) processing the blotch to invert pixel values and perform steps (f) and (g) again;

(i) processing the blotch to increase its contrast and performing steps (f)-(h) again;

(j) processing the blotch to bring the blotch to a more uniform darkness;

(k) processing the uniform darkness blotch to match surroundings in the digital image; and

(l) outputting the restored image for display to a user.

2. The method according to claim 1, wherein processing the blotch to bring the blotch to a uniform darkness includes using a localized value derived from local pixel values and statistical values in the blotch.

3. The method according to claim 1, wherein processing the uniform blotch to match surroundings in the digital image includes a variance parameter, a luminosity parameter, and a tuning parameter.

4. The method according to claim 1, wherein processing the blotch to bring the blotch to a uniform darkness includes computing:

R(ω

_x,y)=L(f,g,h,λ

)

where R(ω

) represents repaired blotch pixels, L is a generic function,

λ

(•

)=Z(i(.),j(.))

f(•

)=Median(I(ω

))

g(•

)=I(ω

_x,y)

h(•

)=2•

Median(I(ω

))

i(•

)=Median(I(ω

))−

I(ω

_x,y)

j(•

)=Median(I(ω

))

5. The method according to claim 1, wherein processing the uniform darkness blotch to match surroundings in the digital image includes computing:

R(ω

)=k(I(ω

))

where k is a generic function,

k(•

)=I(ω

)<sup>α

β

γ</sup>

$\mathrm{\alpha <br><br>}=\frac{1<br><br>(·<br><br>)}{m<br><br>(·<br><br>)}$
l(•

)=log(var[R(ω

)])

m(•

)=log(var[I(ω

)])

$\mathrm{\beta <br><br>}=\frac{n<br><br>(·<br><br>)}{o<br><br>(·<br><br>)}$
n(•

)=log(E[R(ω

)])

o(•

)=log(E[I(ω

)^a]),

where k,α

,l,m,β

, n, and o are generic functions.

6. The method according to claim 1, wherein processing the blotch edge also matches surroundings in the defect and the digital image includes computing:

$R<br><br>\left({\mathrm{\omega <br><br>}}_{x,y}\right)=\left(\frac{p<br><br>(·<br><br>)}{q<br><br>(·<br><br>)}\right)$
p(•

)=ψ

_x,y•

Median[I(ω

)]+ω

_x,y

q(•

)=2

${\mathrm{\psi <br><br>}}_{x,y}={\left(\frac{r<br><br>(·<br><br>)}{s<br><br>(·<br><br>)}\right)}^{{\mathrm{\gamma <br><br>}}_{x,y}}$
r(•

)=ω

_x,y

s(•

)=2•

Median[I(ω

)]

γ

_x,y=t(•

)•

u(•

)

$t<br><br>(·<br><br>)=\frac{v<br><br>(·<br><br>)-{\mathrm{\omega <br><br>}}_{x,y}}{I<br><br>\left(\mathrm{\omega <br><br>}\right)}$$u<br><br>(·<br><br>)=\frac{2}{w<br><br>(·<br><br>)}$
v(•

)=Median[I(ω

)]

w(•

)=StdDev[I(ω

)],

where p, q, ψ

, r, s, γ

, t, u, v, and w generic functions.

7. The method according to claim 1, wherein processing the blotch edge such that the edge also matches its surroundings in both the defect and the digital image includes computing:

R(ω

)=K(ω

_x,y)•

X(•

)

where K(ω

) represents the edge pixels to be restored, and x is a generic function,

x(•

)=exp(γ

•

y(•

))

y(•

)=Median[Λ

]

Λ

=z(•

)

8. The method of claim 1, wherein a set of locations is identified using a separate defect location map such as a mask or any other detection method.

9. The method of claim 1, wherein restoration occurs using local and/or global image processing.

10. The method of claim 1, further including processing an image having a plurality of pixel values for any given pixel location including a color value.

11. The method of claim 1, wherein color planes are separated prior to processing in step (a) and recombined after processing.

12. The method of claim 1, wherein the defect includes a semi-transparent blotch.

13. The method of claim 1, wherein the defect includes text degradation.

14. The method of claim 1, wherein the defect includes a crack.

15. A system, comprising:

a processor and a memory to provide;

an edge analysis module to detect an edge of a defect in a blotch in a digital image;

a uniform intensity module to bring the blotch to a more uniform darkness;

a uniform intensity module to process the more uniform darkness blotch to match surroundings in the digital image;

an edge pixel module to copy an edge of a blotch in the digital image into the processed blotch;

an edge pixel restoration module to process pixels of the edge to repair the copied edge pixels; and

an output module to output the processed image for display to a user.

16. An article, comprising:

a storage medium containing stored instructions that when executed enable a machine to perform;

determining, using a computer having a processor, a plurality of pixel locations relating to a defect and whether a blotch has an edge darker than a selected threshold;

where it is determined that the edge is not darker than the selected threshold,

(a) processing the blotch to bring the blotch to a more uniform darkness;

(b) processing the uniform darkness blotch to match surroundings in the digital image;

(c) copying a border of the blotch in the digital image into the processed blotch;

(d) processing pixels of the border to repair the copied border pixels; and

(e) outputting the processed image for display to a user;

where it is determined that the edge is darker than the selected threshold,

(f) processing the blotch to locally enhance the blotch and match surroundings in the digital image;

(g) processing the blotch edge such that the edge also matches surroundings in the defect and the digital image;

(h) processing the blotch to invert pixel values and perform steps (f) and (g) again;

(i) processing the blotch to increase its contrast and performing steps (f)-(h) again;

(j) processing the blotch to bring the blotch to a more uniform darkness;

(k) processing the uniform darkness blotch to match surroundings in the digital image; and

(l) outputting the restored image for display to a user.