I found this forum and i desperatly need help.We got homework to make some filters but i have never seen all that and i dont know where to start.Could anyone help me and write how to do it?I already have a grade from this but i have to give him this to go on other subjects...Please...
RESTORATION OF DEGRADED IMAGES
* In a system for image acquisition comes often to a lesser or greater damage to the image caused
imperfections of the device which performs the recording. The object of this project is the implementation of
system for the restoration of images degraded by passing through the system shown in Figure 1. description
individual blocks is given below.
Change the size and
geom. degradation
linear
transformation
pixels in HSV
space
Input image Batervortov NF
first-order filter
DCT kvatizacija and
run-length
coding
The output image
Fig.1. Image processing system
* The first block is transformed dimensions of the input image and enter the geometric degradation.
Dimensions of input color images are 512x512 pixels, and after leaving the block dimensions are
640x480 pixels (width x height). Device acquisition brings and geometric degradation. Specifically,
focus image point with coordinates (256.5, 256.5), is mapped to a point with coordinates
(280.5, 180.5), so that the four triangles of the left figure in FIG. 2 affine mapped to
corresponding triangle in the right figure (see cp2tform function in MATLAB).
(256.5, 256.5)
(280.5, 180.5)
*
Fig.2. Transforming the size of the input image and geometric degradation
* In the second part of the system is Hue component images in the HSV color space is replaced
their negatives.
* The third block is Batervortov first-order low-pass filter, which is implemented as a separable filter
(a separabilnih filter applies that H (u, v) = H1 (u) H2 (v)). Limit frequency filters H1 (u) and
H2 (v) amounts uc = 40 in the horizontal direction and vc = 20 in the vertical.
* The last block performs image compression before the final track on the disc. Image compression is performed
in the transformation (DCT) domain. Prior to joining the DCT domain, the image is transferred from the RGB
color space, suitable for compression, YCbCr color space (see function rgb2ycbcr)
is executed and then the decimation factor of 2 and the second picture of a third channel (Cb and Cr). each channel
is divided into 8x8 sub-picture size and is calculated for each DCT transform. The resulting
DCT coefficients are quantized using a quantization matrix given in Tables 1 and 2
Applied as a separate quantization tables for luma (Y) and hrominentne channels (Cb and Cr). SMER of Communication Technologies and Signal Processing
TERMS OF REFERENCE FOR DIGITAL IMAGE PROCESSING (Sk. 2013/14. GOD.)
*
Each 8x8 block konvertivan in a series of 64 quantized DCT coefficients zigzag scanning
(see Table 3). Run-length coding sequence obtained was performed by a prescribed
DC value, and each non-zero AC component is registered as a couple with the number of zeros that her
preceded by a zig-zag line. The last recorded pair (number zero, non-zero AC coefficient) in the block
actually describes a final sequence of zeros that is spread to the end of the block, and therefore in this case of the
Place AC value is a zero, which actually represents the label to arrive at the end of the
block. All run-length encoded DCT coefficient arrays housed in one long string, so
they are first placed run-length arrays of Y channel (their number is 4800), then the Cb channel
(1200 series) and at the end of the Cr channel (1200 series, as well as in Cb). The resulting series is filmed
R as a variable in the MATLAB file projekat.mat. (The variable R is loaded into the MATLAB working
space reference instructions load projekat.mat when this file is located in the working directory
Or simply drag (drag-and-drop) the file projekat.mat the MATLAB workspace.)
Table 1 quantization table for the luma channel (Y) Image
*
*
*
*
*
Table 2 quantization tables hrominentne channels (Cb and Cr) image
*
*
*
*
*
*
Table 3 matrix that defines the zig-zag scan
*
*
*
*
*
16 11 10 16 24 40 51 61
12 12 14 19 26 58 60 55
14 13 16 24 40 57 69 56
14 17 22 29 51 87 80 62
18 22 37 56 68 109 103 77
24 35 55 64 81 104 113 92
49 64 78 87 103 121 120 101
72 92 95 98 112 100 103 99
17 18 24 47 99 99 99 99
18 21 26 66 99 99 99 99
24 26 56 99 99 99 99 99
47 66 99 99 99 99 99 99
99 99 99 99 99 99 99 99
99 99 99 99 99 99 99 99
99 99 99 99 99 99 99 99
99 99 99 99 99 99 99 99
0 1 5 6 14 15 27 28
2 4 7 13 16 26 29 42
3 8 12 17 25 30 41 43
9 11 18 24 31 40 44 53
10 19 23 32 39 45 52 54
20 22 33 38 46 51 55 60
21 34 37 47 50 56 59 61
35 36 48 49 57 58 62 63
RESTORATION OF DEGRADED IMAGES
* In a system for image acquisition comes often to a lesser or greater damage to the image caused
imperfections of the device which performs the recording. The object of this project is the implementation of
system for the restoration of images degraded by passing through the system shown in Figure 1. description
individual blocks is given below.
Change the size and
geom. degradation
linear
transformation
pixels in HSV
space
Input image Batervortov NF
first-order filter
DCT kvatizacija and
run-length
coding
The output image
Fig.1. Image processing system
* The first block is transformed dimensions of the input image and enter the geometric degradation.
Dimensions of input color images are 512x512 pixels, and after leaving the block dimensions are
640x480 pixels (width x height). Device acquisition brings and geometric degradation. Specifically,
focus image point with coordinates (256.5, 256.5), is mapped to a point with coordinates
(280.5, 180.5), so that the four triangles of the left figure in FIG. 2 affine mapped to
corresponding triangle in the right figure (see cp2tform function in MATLAB).
(256.5, 256.5)
(280.5, 180.5)
*
Fig.2. Transforming the size of the input image and geometric degradation
* In the second part of the system is Hue component images in the HSV color space is replaced
their negatives.
* The third block is Batervortov first-order low-pass filter, which is implemented as a separable filter
(a separabilnih filter applies that H (u, v) = H1 (u) H2 (v)). Limit frequency filters H1 (u) and
H2 (v) amounts uc = 40 in the horizontal direction and vc = 20 in the vertical.
* The last block performs image compression before the final track on the disc. Image compression is performed
in the transformation (DCT) domain. Prior to joining the DCT domain, the image is transferred from the RGB
color space, suitable for compression, YCbCr color space (see function rgb2ycbcr)
is executed and then the decimation factor of 2 and the second picture of a third channel (Cb and Cr). each channel
is divided into 8x8 sub-picture size and is calculated for each DCT transform. The resulting
DCT coefficients are quantized using a quantization matrix given in Tables 1 and 2
Applied as a separate quantization tables for luma (Y) and hrominentne channels (Cb and Cr). SMER of Communication Technologies and Signal Processing
TERMS OF REFERENCE FOR DIGITAL IMAGE PROCESSING (Sk. 2013/14. GOD.)
*
Each 8x8 block konvertivan in a series of 64 quantized DCT coefficients zigzag scanning
(see Table 3). Run-length coding sequence obtained was performed by a prescribed
DC value, and each non-zero AC component is registered as a couple with the number of zeros that her
preceded by a zig-zag line. The last recorded pair (number zero, non-zero AC coefficient) in the block
actually describes a final sequence of zeros that is spread to the end of the block, and therefore in this case of the
Place AC value is a zero, which actually represents the label to arrive at the end of the
block. All run-length encoded DCT coefficient arrays housed in one long string, so
they are first placed run-length arrays of Y channel (their number is 4800), then the Cb channel
(1200 series) and at the end of the Cr channel (1200 series, as well as in Cb). The resulting series is filmed
R as a variable in the MATLAB file projekat.mat. (The variable R is loaded into the MATLAB working
space reference instructions load projekat.mat when this file is located in the working directory
Or simply drag (drag-and-drop) the file projekat.mat the MATLAB workspace.)
Table 1 quantization table for the luma channel (Y) Image
*
*
*
*
*
Table 2 quantization tables hrominentne channels (Cb and Cr) image
*
*
*
*
*
*
Table 3 matrix that defines the zig-zag scan
*
*
*
*
*
16 11 10 16 24 40 51 61
12 12 14 19 26 58 60 55
14 13 16 24 40 57 69 56
14 17 22 29 51 87 80 62
18 22 37 56 68 109 103 77
24 35 55 64 81 104 113 92
49 64 78 87 103 121 120 101
72 92 95 98 112 100 103 99
17 18 24 47 99 99 99 99
18 21 26 66 99 99 99 99
24 26 56 99 99 99 99 99
47 66 99 99 99 99 99 99
99 99 99 99 99 99 99 99
99 99 99 99 99 99 99 99
99 99 99 99 99 99 99 99
99 99 99 99 99 99 99 99
0 1 5 6 14 15 27 28
2 4 7 13 16 26 29 42
3 8 12 17 25 30 41 43
9 11 18 24 31 40 44 53
10 19 23 32 39 45 52 54
20 22 33 38 46 51 55 60
21 34 37 47 50 56 59 61
35 36 48 49 57 58 62 63
