Calculation of degree of similarity between two images
An image processing apparatus includes: a first circuit which calculates values f(R Pi ), f(G Pi ) and f(B Pi ) by applying a function f(x) to an R grayscale value R Pi , a G grayscale value G Pi and a B grayscale value B Pi of each pixel i of a first image; a second circuit which calculates values f(R Qi ), f(G Qi ) and f(B Qi ) by applying the function f(x) to an R grayscale value R Qi , a G grayscale value G Qi and a B grayscale value B Qi of each pixel i of a second image; and a similarity calculation circuit which calculates a degree of similarity between the first and second images depending on |f(R Pi )−f(R Qi )|, |f(G Pi )−|f(G Qi )| and |f(B Pi )−f(B Qi )| associated with each pixel i of the first and second images. The function f(x) is a convex function monotonically non-decreasing in the domain of definition.
1. An image processing apparatus, comprising:
a first circuit operable to:
receive RGB image data of a first image;
for each pixel i of the first image, apply a function f(x) to an R grayscale value R Pi , a G grayscale value G Pi , and a B grayscale value B Pi of the pixel; and
output values f(R Pi ), f(G Pi ), and f(B Pi );
a second circuit operable to:
receive RGB image data of a second image;
for each pixel i of the second image, apply the function f(x) to an R grayscale value R Qi , a G grayscale value G Qi , and a B grayscale value B Qi of the pixel; and
output values f(R Qi ), f(G Qi ), and f(B Qi );
a similarity calculation circuit operable to:
receive the values f(R Pi ), f(G Pi ), f(B Pi ), f(R Qi ), f(G Qi ), and f(B Qi ); and
output a similarity value representing a degree of similarity between the first image and the second image depending on |f(R Pi )−f(R Qi )|, |f(G Pi )−f(G Qi )|, and |f(B Pi )−f(B Qi )| of each pixel i of the first image and the second image; and
a third circuit operable to:
perform, based on the similarity value, an image processing operation on the RGB image data of the first image,
wherein a lower limit of a domain of definition of the function f(x) is an allowed minimum value of the R grayscale values R Pi , R Qi , the G grayscale values G Pi , G Qi , and the B grayscale values B Pi , B Qi ,
wherein an upper limit of the domain of definition is an allowed maximum value of the R grayscale values R Pi , R Qi , the G grayscale values G Pi , G Qi , and the B grayscale values B Pi , B Qi , and
wherein the function f(x) is a convex function monotonically non-decreasing in the domain of definition.
2. The image processing apparatus according to claim 1 , wherein the degree of similarity between the first and second images is calculated in accordance with the following expression (1):
S
=
∑
i
g
(
|
f
(
R
Pi
)
-
f
(
R
Qi
)
|
,
|
f
(
G
Pi
)
-
f
(
G
Qi
)
|
,
|
f
(
B
Pi
)
-
f
(
B
Qi
)
|
)
,
(
1
)
where Σ represents a sum with respect to all pixels of the first and second images and g(x, y, z) is a function which is not a constant function.
3. The image processing apparatus according to claim 2 , further comprising:
a second setting register storing a second setting parameter specifying a coefficient included in the function g(x, y, z).
4. The image processing apparatus according to claim 1 , wherein the degree of similarity between the first and second images is calculated in accordance with the following expression (2):
S
=
∑
i
{
K
R
|
f
(
R
Pi
)
-
f
(
R
Qi
)
|
p
+
K
G
|
f
(
G
Pi
)
-
f
(
G
Qi
)
|
p
+
K
B
|
f
(
B
Pi
)
-
f
(
B
Qi
)
|
p
}
,
(
2
)
wherein p is a non-zero number and K R , K G and K B are weighting factors defined for red, green and blue colors, respectively, in calculating the degree of similarity S.
5. The image processing apparatus according to claim 1 , wherein the degree of similarity between the first and second images is calculated in accordance with the following expression (3):
S
=
∑
i
{
K
R
|
f
(
R
Pi
)
-
f
(
R
Qi
)
|
+
K
G
|
f
(
G
Pi
)
-
f
(
G
Qi
)
|
+
K
B
|
f
(
B
Pi
)
-
f
(
B
Qi
)
|
}
,
(
3
)
wherein K R , K G and K B are weighting factors defined for red, green and blue colors, respectively, in calculating the degree of similarity S.
6. The image processing apparatus according to claim 5 , wherein the weighting factors K R , K G and K B are determined based on a color gamut of a display panel included in a display device in which image data of the first and second images are used.
7. The image processing apparatus according to claim 5 , further comprising:
a second setting register storing second setting parameters specifying the weighting factors K R , K G and K B .
8. The image processing apparatus according to claim 1 , wherein the degree of similarity between the first and second images is calculated in accordance with the following expression (4):
S
=
∑
i
{
K
R
|
R
Pi
γ
-
R
Qi
γ
|
+
K
G
|
G
Pi
γ
-
G
Qi
γ
|
+
K
B
|
B
Pi
γ
-
B
Qi
γ
|
}
,
(
4
)
wherein K R , K G and K B are weighting factors defined for red, green and blue colors, respectively, in calculating the degree of similarity S and γ is a predetermined constant.
9. The image processing apparatus according to claim 1 , wherein the function f(x) is defined as a polygonal line function which approximates x^γ, where γ is a predetermined constant.
10. The image processing apparatus according to claim 9 , wherein γ is equal to a gamma value of a display device in which image data of the first and second images.
11. The image processing apparatus according to claim 1 , wherein the function f(x) is defined as a polynomial function which approximates x^γ, where γ is a predetermined constant.
12. The image processing apparatus according to claim 1 , further comprising:
a first setting register storing a first setting parameter specifying a coefficient included in the function f(x).
13. The image processing apparatus of claim 1 , wherein performing the image processing operation on the RGB image data of the first image comprises:
selecting, based on the similarity value, a first compression process of a plurality of compression processes; and
outputting compressed image data that is compressed according to the first compression process.
14. An image compression circuit, comprising:
a compression circuity generating first to N th compressed data by performing first to N th compression processes on an original image data, respectively, for N being an integer of two or more;
a decompression circuitry generating first to N th decompressed data by respectively performing corresponding decompression processes on the first to N th compressed data;
first to (N+1) th grayscale data conversion circuits; and
a compressed image data selection circuit selecting an output compressed image data from among the first to N th compressed data and outputting the output compressed image data,
wherein the k th grayscale data conversion circuit of the first to N th grayscale data conversion circuit is configured to calculate values f(R ki ), f(G ki ) and f(B ki ) by respectively applying a function f(x) to an R grayscale value R ki , a G grayscale value G ki and a B grayscale value B ki of each pixel i of the k th decompressed data of the first to N th decompressed data, for k being any integer from one to N,
wherein the (N+1) th grayscale data conversion circuit is configured to calculate values f(R INi ), f(G INi ) and f(B INi ) by respectively applying the function f(x) to an R grayscale value R INi , a G grayscale value G INi and a B grayscale value B INi of each pixel i of the original image data,
wherein the compressed image data select circuit is configured to calculate degrees of similarity between the original image data and the first to N th decompressed data and select an output compressed image data from the first to N th compressed data in response to the calculated degrees of similarity,
wherein the degree of similarity between the k th decompressed data and the original image data is calculated depending on |f(R ki )−f(R INi )|, |f(G ki )−f(G INi )| and |f(B ki )−f(B INi )| associated with each pixel i of the k th decompressed data and the original image data,
wherein a lower limit of a domain of definition of the function f(x) is an allowed minimum value of the R grayscale values R ki , R INi , the G grayscale values G ki , G INi and the B grayscale values B ki , B INi ,
wherein an upper limit of the domain of definition is an allowed maximum value of the R grayscale values R ki , R INi , the G grayscale values G ki , G INi and the B grayscale values B ki B INi , and
wherein the function f(x) is a convex function monotonically non-decreasing in the domain of definition.
15. The compression circuit according to claim 14 , wherein the degree of similarity S k between the k th decompressed data and the original image data is calculated in accordance with the following expression (5):
S
k
=
∑
i
{
K
R
|
f
(
R
ki
)
-
f
(
R
INi
)
|
+
K
G
|
f
(
G
ki
)
-
f
(
G
INi
)
|
+
K
B
|
f
(
B
ki
)
-
f
(
B
INi
)
|
}
,
(
5
)
where K R , K G and K B are weighting factors defined for red, green and blue colors, respectively, in calculating the degree of similarity S k .
16. The compression circuit according to claim 14 , wherein the degree of similarity S k between the k th decompressed data and the original image data is calculated in accordance with the following expression (6):
S
k
=
∑
i
{
K
R
|
R
ki
γ
-
R
INi
γ
|
+
K
G
|
G
ki
γ
-
G
INi
γ
|
+
K
B
|
B
ki
γ
-
B
INi
γ
|
}
,
(
6
)
where K R , K G and K B are weighting factors defined for red, green and blue colors, respectively, in calculating the degree of similarity S k and γ is a predetermined constant.
17. The compression circuit according to claim 14 , wherein the function f(x) is defined as a polygonal line function which approximates x^γ, where γ is a predetermined constant.
18. A display driver for driving a display panel, comprising:
an image compression circuit generating an output compressed image data from an original image data;
an image memory storing the output compressed image data received from the compression circuit;
an image decompression circuit generating an decompressed image data by decompressing the output compressed image data received from the image memory; and
a drive circuitry driving the display panel in response to the decompressed image data,
wherein the image compression circuit includes:
a compression circuity generating first to N th compressed data by performing first to N th compression processes on the original image data, respectively, for N being an integer of two or more;
a decompression circuitry generating first to N th decompressed data by respectively performing corresponding decompression processes on the first to N th compressed data;
first to (N+1) th grayscale data conversion circuits; and
a compressed image data selection circuit selecting the output compressed image data from among the first to N th compressed data and outputting the output compressed image data,
wherein the k th grayscale data conversion circuit of the first to N th grayscale data conversion circuit is configured to calculate values f(R ki ), f(G ki ) and f(B ki ) by respectively applying a function f(x) to an R grayscale value R ki , a G grayscale value G ki and a B grayscale value B ki of each pixel i of the k th decompressed data of the first to N th decompressed data, for k being any integer from one to N,
wherein the (N+1) th grayscale data conversion circuit is configured to calculate values f(R INi ), f(G INi ) and f(B INi ) by respectively applying the function f(x) to an R grayscale value R INi , a G grayscale value G INi and a B grayscale value B INi of each pixel i of the original image data,
wherein the compressed image data select circuit is configured to calculate degrees of similarity between the original image data and the first to N th decompressed data and select an output compressed image data from the first to N th compressed data in response to the calculated degrees of similarity,
wherein the degree of similarity between the k th decompressed data and the original image data is calculated depending on |f(R Pi )−f(R Qi )|, |f(G Pi )−f(G Qi )| and |f(B Pi )−f(B Qi )| associated with each pixel i of the k th decompressed data and the original image data,
wherein a lower limit of a domain of definition of the function f(x) is an allowed minimum value of the R grayscale values R ki , R INi , the G grayscale values G ki , G INi and the B grayscale values B ki , B INi , and
wherein an upper limit of the domain of definition is an allowed maximum value of the R grayscale values R ki , R INi , the G grayscale values G ki , G INi and the B grayscale values B ki , B INi , and
wherein the function f(x) is a convex function monotonically non-decreasing in the domain of definition.
19. The display driver according to claim 18 , further comprising:
a first setting register storing a first setting parameter specifying a coefficient included in the function f(x).
20. The display driver according to claim 18 , further comprising a second setting register,
wherein the degree of similarity S k between the k th decompressed image data and the original image data is calculated in accordance with the following expression (7):
S
k
=
∑
i
{
K
R
|
f
(
R
ki
)
-
f
(
R
INi
)
|
+
K
G
|
f
(
G
ki
)
-
f
(
G
INi
)
|
+
K
B
|
f
(
B
ki
)
-
f
(
B
INi
)
|
}
,
(
7
)
where K R , K G and K B are weighting factors defined for red, green and blue colors, respectively, in calculating the degree of similarity S k , and
wherein the second setting register stores second setting parameters specifying the weighting factors K R , K G and K B .
21. The display driver according to claim 18 ,
wherein the degree of similarity S k between the k th decompressed image data and the original image data is calculated in accordance with the following expression (8):
S
k
=
∑
i
{
K
R
|
R
Pi
γ
-
R
Qi
γ
|
+
K
G
|
G
Pi
γ
-
G
Qi
γ
|
+
K
B
|
B
Pi
γ
-
B
Qi
γ
|
}
,
(
8
)
where K R , K G and K B are weighting factors defined for red, green and blue colors, respectively, in calculating the degree of similarity S k ,
wherein γ is equal to a gamma value of the display panel.