Granted Patent B2

US 9,241,093 · App. 14/220,528 · Granted Jan 19, 2016

Abstract

Claims (170)

1. A system for generating a detail-enhanced video signal, the system comprising;

a filter that receives an input video signal, the filter being configured to:

generate a first signal that is a linear transformation of the input video signal, and

generate a detail signal by determining a difference between the input video signal and the first signal, wherein the detail signal includes information that is added to the input video signal to generate the detail-enhanced video signal;

a noise reducer configured to generate a filtered detail signal by removing noise from the detail signal;

a multiplier configured to multiply the filtered detail signal by a gain function to generate a second signal, the gain function being based on an amount of variance in the input video signal; and

an adder configured to combine the input video signal and the second signal to generate the detail-enhanced video signal.

2. The system of claim 1 ,

wherein the input video signal comprises a first plurality of pixels;

wherein the filtered detail signal comprises a second plurality of pixels;

wherein the multiplier is configured to multiply a pixel of the second plurality of pixels and a gain value of the gain function;

wherein the gain value is based on a variance of a group of pixels that includes a pixel of the first plurality of pixels; and

wherein the pixel of the first plurality of pixels and the pixel of the second plurality of pixels have same x-y coordinates.

3. The system of claim 1 , wherein the second signal is defined by

Δ Y det =( D −noise)*gain,

where ΔY det is the second signal, D is the detail signal, noise is the noise removed from the detail signal, and gain is the gain function.

4. The system of claim 1 , wherein the linear transformation is defined by

q i =ā i I i + b i ,

where I i is a value of a pixel i of the input video signal, q i is a pixel of the first signal corresponding to the pixel i, ā is a first average coefficient that is indicative of the variance of a group of pixels including the pixel i, b is a second average coefficient that is indicative of a mean value of the group of pixels.

5. The system of claim 4 , wherein the gain function is based on the variable ā.

6. The system of claim 4 , wherein the wain function is defined by

gain

=

{

scale

*

exp

⁡

(

-

(

a

_

-

mid

)

2

σ

g

⁢

⁢

1

2

)

for

⁢

⁢

a

_

<

mid

scale

*

exp

⁡

(

-

(

a

_

-

mid

)

2

σ

g

⁢

⁢

2

2

)

for

⁢

⁢

a

_

>=

mid

,

where gain is a value of the gain function at a particular value of ā, and mid, σ g1 2 , σ g2 2 , and scale are parameters that control a distribution of the gain function.

7. The system of claim 4 , wherein the noise reducer utilizes an equation for calculating the noise in the detail signal, and wherein the equation is based on the variable ā.

8. A method for generating a detail-enhanced video signal, the method comprising:

receiving an input video signal;

generating a first signal that is a linear transformation of the input video signal;

generating a detail signal by determining a difference between the input video signal and the first signal, wherein the detail signal includes information that is added to the input video signal to generate the detail-enhanced video signal;

generating a filtered detail signal by removing noise from the detail signal;

multiplying the filtered detail signal by a gain function to generate a second signal, the gain function being based on an amount of variance in the input video signal; and

combining the input video signal and the second signal to generate the detail-enhanced video signal.

9. The method of claim 8 , wherein the input video signal comprises a first plurality of pixels, and wherein the filtered detail signal comprises a second plurality of pixels, the method further comprising:

multiplying a pixel of the second plurality of pixels and a gain value of the gain function, wherein the gain value is based on a variance of a group of pixels that includes a pixel of the first plurality of pixels, and wherein the pixel of the first plurality of pixels and the pixel of the second plurality of pixels have same x-y coordinates.

10. The method of claim 8 , wherein the second signal is defined by

Δ Y det =( D −noise)*gain,

where ΔY det is the second signal. D is the detail signal, noise is the noise removed from the detail signal, and gain is the gain function.

11. The method of claim 8 . wherein the linear transformation is defined by

q i =ā i I i + b i ,

where I i is a value of a pixel i of the input video signal, q i is a pixel of the first signal corresponding to the pixel i, ā is a first average coefficient that is indicative of the variance of a group of pixels including the pixel i, b is a second average coefficient that is indicative of a mean value of the group of pixels.

12. The method of claim 11 , wherein the gain function is based on the variable ā.

13. The method of claim 11 , wherein the gain function is defined by

gain

=

{

scale

*

exp

⁡

(

-

(

a

_

-

mid

)

2

σ

g

⁢

⁢

1

2

)

for

⁢

⁢

a

_

<

mid

scale

*

exp

⁡

(

-

(

a

_

-

mid

)

2

σ

g

⁢

⁢

2

2

)

for

⁢

⁢

a

_

>=

mid

,

where gain is a value of the gain function at a particular value of ā, and mid, σ g1 2 , σ g2 2 , and scale are parameters that control a distribution of the vain function.

14. The method of claim 11 , further comprising:

utilizing an equation for calculating the noise in the detail signal, wherein the equation is based on the variable ā.

15. The system of claim 1 , wherein the multiplying, by the multiplier includes, for each respective pixel in the filtered detail signal, multiplying the respective pixel by the gain value to generate a correspondingly-located pixel in the second signal, and

wherein the gain value is a function of variance in a region that the respective pixel is located in, such that (i) the gain value peaks at a mid value of variance, (ii) from the mid value the gain value decreases with increasing variance, and (iii) from the mid value the gain value decreases with decreasing variance.

16. The system of claim 15 , wherein the function of the gain value versus variance is bowed from a variance value below the mid value to a variance value above the mid value.

17. The method of claim 8 , wherein the multiplying includes, for each respective pixel in the filtered detail signal, multiplying the respective pixel by the gain value to generate a correspondingly-located pixel in the second signal, and

wherein the gain value is a function of variance in a region that the respective pixel is located in, such that (i) the gain value peaks at a mid value of variance, (ii) from the mid value the gain value decreases with increasing variance, and. (iii) from the mid value the gain value decreases with decreasing variance.

18. The method of claim 17 , wherein the function of the gain value versus variance is bowed from a variance value below the mid value to a variance value above the mid value.

19. The system of claim 1 , wherein the system is configured, for a pixel that is in the input signal and that has a color within the ranges 0°≦H≦50° and 0.2≦S≦0.68 and 0.35≦V≦1, to (i) set the gain function for the pixel to a gain value that is predetermined for skin and (ii) set the noise to be removed for the pixel to a noise value that is predetermined for skin.

20. The method of claim 8 , further comprising

for a pixel, in the input signal, having a color within the ranges 0°≦H≦50° and 0.2≦S≦0.68 and 0.35≦V≦1, both

setting the gain function for the pixel to a gain value that is predetermined for skin, and

setting the noise to be removed for the pixel to a noise value that is predetermined for skin.

Assignments (7)