Method for accessing image data and method for accessing YUV420 image data
A method for accessing image data and a method for accessing YUV420 image data are disclosed. The method for accessing image data uses two different approaches to a buffer for accessing the JPEG image data, so that the method can achieve instant JPEG image data compression with only one buffer.
1. A method for accessing image data, the image data comprising a plurality of image data matrixes A, each said image data matrix A comprising m rows of image data sub-matrixes R, each said image data sub-matrix R comprising n columns of image units a, each said image unit a comprising p pixels, each said pixel being represented with a plurality of units of data, the plurality of units comprising a first unit, a second unit and a third unit, wherein data of the first unit of w rows of image data sub-matrixes require u rows, each row comprising n columns, of memory units b for storage, and data of the second unit and the third unit of the w rows of image data sub-matrixes requires v rows, each row comprising n columns of the memory units b for storage, wherein
A
=
[
R
1
R
2
⋮
R
m
]
=
[
a
1
,
1
a
1
,
2
…
a
1
,
n
a
2
,
1
a
2
,
2
…
a
2
,
n
⋮
⋮
⋮
a
m
,
1
a
m
,
2
…
a
m
,
n
]
,
wherein R i represents the i th row of image data sub-matrix R; a i,j represents the i th row and j th column image unit a, the method for accessing image data comprising the following steps:
(a) providing a matrix storage space B, for temporarily storing the image data matrixes A, comprising x rows of sub-matrix storage spaces S or z columns of sub-matrix storage spaces C, each said sub-matrix storage space S comprising y columns of the memory units b, each said sub-matrix storage space C comprising x rows of sub-matrix storage spaces T or h columns of sub-matrix storage spaces D, each said sub-matrix storage space T comprising h columns of memory units b, wherein x=m*(u+v)/w, and when n*(1+v/u)/x is an integer, h=n*(1+v/u)/x, and when n*(1+v/u)/x is not an integer, h=INT(n*(1+v/u)/x)+1, wherein the function INT( ) is adapted to obtain an integer portion of an input parameter, and when n/h is an integer, z=n/h, and when n/h is not an integer, z=INT(n/h)+1, and y=h*z, wherein
B
=
[
S
1
S
2
⋮
S
x
]
=
[
b
1
,
1
b
1
,
2
…
b
1
,
y
b
2
,
1
b
2
,
2
…
b
2
,
y
⋮
⋮
⋮
b
x
,
1
b
x
,
2
…
b
x
,
y
]
=
[
C
1
C
2
…
C
z
]
,
C
q
=
[
T
1
T
2
⋮
T
x
]
=
[
b
1
,
1
b
1
,
2
…
b
1
,
h
b
2
,
1
b
2
,
2
…
b
2
,
h
⋮
⋮
⋮
b
x
,
1
b
x
,
2
…
b
x
,
h
]
=
[
D
1
D
2
…
D
h
]
,
wherein S k represents the kth row of sub-matrix storage spaces S, b k,l represents the k th row and l th column memory unit b, C q represents the q th column of sub-matrix storage spaces C, and T k represents the k th row of sub-matrix storage spaces T;
(b) obtaining data of w rows of image data sub-matrixes R in the e th image data matrix A, and writing the data into (u+v) rows of the sub-matrix storage spaces S of the matrix storage space B;
(c) reading data of the e th image data matrix A from the sub-matrix storage space C of the matrix storage space B, wherein after data of the one column of sub-matrix storage spaces C are read out completely, data of another column of sub-matrix storage spaces C can be read;
(d) obtaining data of w rows of the image data sub-matrixes R in the (e+1) th image data matrix A, and writing the data into the sub-matrix storage spaces C of the matrix storage space B; and
(e) reading data of the (e+1) th image data matrix A from the (u+v) rows of sub-matrix storage space S of the matrix storage space B, wherein after data of the one group of (u+v) rows of sub-matrix storage space S are read out completely, data of another group of (u+v) rows of sub-matrix storage space S can be read,
wherein e, h, I, j, k, l, m, n, p, q, x, y, and z are natural numbers; w, u, v are rational numbers; and i≦m, j≦n, k≦x, l≦y, and q≦z.
2. The method for accessing image data according to claim 1 , wherein when writing data into the (u+v) rows of sub-matrix storage spaces S of the matrix storage space B, a sequence of writing the data into the memory unit b k,l increases from l=1 to l=n, and when reading the data from the sub-matrix storage spaces C of the matrix storage space B, the data are read from the sub-matrix storage spaces D.
3. The method for accessing image data according to claim 2 , wherein when writing data into the (u+v) rows of sub-matrix storage spaces S of the matrix storage space B, (u−v) rows of data of the first unit, and v rows of data of the second unit and the third unit are written starting from the k th row of sub-matrix storage spaces S k , wherein k increases from k=1 to k=f, and v rows of data of the first unit are written starting from the g th row of sub-matrix storage spaces S g , wherein g increases from g=f+1 to g=x, wherein f=x*u/(u+v), and f and g are both natural numbers.
4. The method for accessing image data according to claim 2 , wherein when writing data into the (u+v) rows of sub-matrix storage spaces S of the matrix storage space B, data of the first unit are written starting from the k th row of sub-matrix storage spaces S k , wherein k increases from k=1 to k=f, and data of the second unit and the third unit are written starting from the g th row of sub-matrix storage spaces S g , wherein g increases from g=f+1 to g=x, wherein f=x*u/(u+v), and f and g are both natural numbers.
5. The method for accessing image data according to claim 1 , wherein when writing the data into the sub-matrix storage spaces C of the matrix storage space B, a sequence of writing the data into the memory units b k,l first increases l from l=1 to l=h, then increases k, and when reading data from (u+v) rows of sub-matrix storage spaces S of the matrix storage space B, the data are read from corresponding memory units b k,l of the sub-matrix storage spaces C, and the reading sequence of the sub-matrix storage spaces C q increases q from q=1 to q=z.
6. The method for accessing image data according to claim 1 , wherein p=8, m=16, n=80, w=2, u=1, v=½, x=12 and y=80.
7. The method for accessing image data according to claim 1 , wherein the data of the first unit are Y image component data, the data of the second unit are U image component data, and the data of the third unit are V image component data.
8. A method for accessing YUV420 image data, the YUV420 image data comprising a plurality of image data matrixes A, each said image data matrix A comprising m rows of image data sub-matrixes R, each said image data sub-matrix R comprising n columns of image units a, each said image unit a comprising p pixels, each pixel being represented with a plurality of units of data, the plurality of units comprising a first unit, a second unit and a third unit, wherein data of the first unit of w rows of image data sub-matrixes require u rows of memory units b for storage, and data of the second unit and the third unit of the w rows of image data sub-matrixes requires v rows of the memory units b for storage, wherein
A
=
[
R
1
R
2
⋮
R
m
]
=
[
a
1
,
1
a
1
,
2
…
a
1
,
n
a
2
,
1
a
2
,
2
…
a
2
,
n
⋮
⋮
⋮
a
m
,
1
a
m
,
2
…
a
m
,
n
]
,
wherein R i represents the i th row of image data sub-matrix R; a i,j represents the i th row and j th column image unit a, the method for accessing YUV420 image data comprising the following steps:
(f) providing a matrix storage space B, for temporarily storing the image data matrixes A, the matrix storage space B comprising x rows of sub-matrix storage spaces S or z columns of sub-matrix storage spaces C, each said sub-matrix storage space S comprising y columns of the memory units b, each said sub-matrix storage space C comprising x rows of sub-matrix storage spaces T or h columns of sub-matrix storage spaces D, each said sub-matrix storage space T comprising h columns of memory units b, each said matrix storage space C further comprising Y rows of sub-matrix storage spaces E, and each said sub-matrix storage space E comprising three rows of sub-matrix storage spaces T, wherein x=m*(u+v)/w, and when n*(1+v/u)/x is an integer, h=n*(1+v/u)/x, and when n*(1+v/u)/x is not an integer, h=INT(n*(1+v/u)/x)+1, wherein the function INT( ) is adapted to obtain an integer portion of an input parameter, and when n/h is an integer, z=n/h, when n/h is not an integer, z=INT(n/h)+1, and y=h*z, when an arithmetical compliment of (2*h)/3Q is zero, then m 1 =Q and m 2 =Q; when an arithmetical compliment of (2*h)/3Q is one, then m 1 =Q and m 2 =Q+1; and when an arithmetical compliment of (2*h)/3Q is two, then m 1 =Q+1 and m=Q, wherein Q is a quotient of (2*h)/3, and when an arithmetical compliment of x/3 is zero, then Y=x/3, wherein
B
=
[
S
1
S
2
⋮
S
x
]
=
[
b
1
,
1
b
1
,
2
…
b
1
,
y
b
2
,
1
b
2
,
2
…
b
2
,
y
⋮
⋮
⋮
b
x
,
1
b
x
,
2
…
b
x
,
y
]
=
[
C
1
C
2
…
C
z
]
,
C
q
=
[
T
1
T
2
⋮
T
x
]
=
[
b
1
,
1
b
1
,
2
…
…
b
1
,
m
1
b
1
,
m
1
+
1
…
b
1
,
h
b
2
,
1
b
2
,
2
…
b
2
,
m
2
b
2
,
m
2
+
1
…
…
b
2
,
h
b
3
,
1
b
3
,
2
…
…
b
3
,
m
1
b
3
,
m
1
+
1
…
b
3
,
h
b
4
,
1
b
4
,
2
…
…
b
4
,
m
1
b
4
,
m
1
+
1
…
b
4
,
h
b
5
,
1
b
5
,
2
…
b
5
,
m
2
b
5
,
m
2
+
1
…
…
b
5
,
h
b
6
,
1
b
6
,
2
…
…
b
6
,
m
1
b
6
,
m
1
+
1
…
b
6
,
h
⋮
⋮
⋮
⋮
⋮
⋮
⋮
⋮
b
x
-
2
,
1
b
x
-
2
,
2
…
…
b
x
-
2
,
m
1
b
x
-
2
,
m
1
+
1
…
b
x
-
2
,
h
b
x
-
1
,
1
b
x
-
1
,
2
…
b
x
-
1
,
m
2
b
x
-
1
,
m
2
+
1
…
…
b
x
-
1
,
h
b
x
,
1
b
x
,
2
…
…
b
x
,
m
1
b
x
,
m
1
+
1
…
b
x
,
h
]
=
[
D
1
D
2
…
D
h
]
=
[
E
1
E
2
⋮
E
Y
]
,
E
k
=
[
T
k
-
1
T
k
T
k
+
1
]
,
wherein S k represents kth sub-matrix storage spaces S, b k,l represents the k th row and l th column memory unit b, C q represents the q th column of sub-matrix storage spaces C, D q represents the q th column of sub-matrix storage spaces D; T k represents the k th row of sub-matrix storage spaces T; and E k represents the k th row of sub-matrix storage spaces E;
(g) obtaining data of w rows of image data sub-matrixes R in the e th image data matrix A, and writing data into (u+v) rows of the sub-matrix storage spaces S of the matrix storage space B;
(h) reading data of the e th image data matrix A from the sub-matrix storage space C of the matrix storage space B, wherein after data of the one column of sub-matrix storage spaces C are read out completely, data of another column of sub-matrix storage spaces C can be read;
(i) obtaining data of w rows of image data sub-matrixes R in the (e+1) th image data matrix A, and writing the data into the sub-matrix storage spaces C of the matrix storage space B; and
(j) reading data of the (e+1) th image data matrix A from the sub-matrix storage space S of the matrix storage space B,
wherein e, h, I, j, k, l, m, n, p, q, x, y, z, m 1 , m 2 , and Q are natural numbers;
w, u, v are rational numbers; and i≦m, j≦n, k≦x, l≦y, and q≦z.
9. The method for accessing YUV420 image data according to claim 8 , wherein when writing data into the (u+v) rows of sub-matrix storage spaces S of the matrix storage space B, a sequence of writing the data into the memory unit bk,l increases from l=1 to l=n, and when reading the data from the sub-matrix storage spaces C of the matrix storage space B, the data are read from the sub-matrix storage spaces D.
10. The method for accessing YUV420 image data according to claim 9 , wherein when writing data into the (u+v) rows of sub-matrix storage spaces S of the matrix storage space B, (u−v) rows of data of the first unit, and v rows of data of the second unit and the third unit are written starting from the k th row of sub-matrix storage spaces S k , wherein k increases from k=1 to k=f, and v rows of data of the first unit are written starting from the g th row of sub-matrix storage spaces S g , wherein g increases from g=f+1 to g=x, wherein f=x*u/(u+v), and f and g are both natural numbers.
11. The method for accessing YUV420 image data according to claim 9 , wherein when writing data into the (u+v) rows of sub-matrix storage spaces S of the matrix storage space B, data of the first unit are written starting from the kth row of sub-matrix storage spaces S k , wherein k increases from k=1 to k=f, and data of the second unit and the third unit are written starting from the g th row of sub-matrix storage spaces S g , wherein g increases from g=f+1 to g=x, wherein f=x*u/(u+v), and f and g are both natural numbers.
12. The method for accessing YUV420 image data according to claim 8 , wherein when writing the data into the sub-matrix storage spaces C of the matrix storage space B, the (u−v) rows of data of the first unit and the v rows of data of the second and the third units are written into the memory units b k−1,l through b k−1,m1 of the sub-matrix storage space T k−1 , b k,l through b k,m2 of the sub-matrix storage space T k , and b k+1,l through b k+1,m1 of the sub-matrix storage space T k+1 of the sub-matrix storage space E.
13. The method for accessing YUV420 image data according to claim 12 , wherein when writing the data into the sub-matrix storage spaces C of the matrix storage space B, the v rows of data of the first unit are written into the memory units b k−1,m1+1 through b k−1,h of the sub-matrix storage space T k−1 , b k,m2+1 through b k,h of the sub-matrix storage space T k , and b k+1,m1+1 through b k+1,h of the sub-matrix storage space T k+1 of the sub-matrix storage space E.
14. The method for accessing YUV420 image data according to claim 8 , wherein p=8, m=16, n=80, w=2, u=1, v=½, x=12 and y=80.
15. The method for accessing YUV420 image data according to claim 14 , wherein when writing data into the (u+v) rows of sub-matrix storage spaces S of the matrix storage space B, the (u−v) rows of data of the first unit and the v rows of data of the second and the third units are written from the k th row of sub-matrix storage spaces S k , wherein k is 1, 3, 4, 6, 7, 9, 10, and 12, and the v rows of data of the first unit are written from the g th row of sub-matrix storage spaces S g , wherein g is 2, 5, 8, and 11.
16. The method for accessing YUV420 image data according to claim 8 , wherein the data of the first unit are Y image component data, the data of the second unit are U image component data, and the data of the third unit are V image component data.