Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd Logo
Upload

Welcome to Scribd!

  • 75 views

    DSP Lab 13

    Uploaded by

    mpssassygirl
    The document describes MATLAB programs for grayscale and color JPEG image compression. For grayscale images, the program divides the image into 8x8 blocks, applies discrete cosine transform (DCT) to each block, quantizes the DCT coefficients using a quantization matrix, and reconstructs the image. Reducing the quantization factor increases compression but decreases quality. For color images, the program converts RGB to YIQ components, applies separate DCT and quantization to each channel, then reconstructs the RGB image. Running the programs with different quantization factors for Y, I, and Q channels affects the range of quantized DCT coefficients and perceived image quality.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as DOC, PDF, TXT or read online from Scribd

    DSP Lab 13

    Uploaded by

    mpssassygirl
    75 views5 pages
    The document describes MATLAB programs for grayscale and color JPEG image compression. For grayscale images, the program divides the image into 8x8 blocks, applies discrete cosine transform (DCT) to each block, quantizes the DCT coefficients using a quantization matrix, and reconstructs the image. Reducing the quantization factor increases compression but decreases quality. For color images, the program converts RGB to YIQ components, applies separate DCT and quantization to each channel, then reconstructs the RGB image. Running the programs with different quantization factors for Y, I, and Q channels affects the range of quantized DCT coefficients and perceived image quality.

    Original Description:

    Demonstration of DCT algorithms using matlab

    Original Title

    DSP_Lab_13

    Copyright

    © Attribution Non-Commercial (BY-NC)

    Available Formats

    DOC, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document describes MATLAB programs for grayscale and color JPEG image compression. For grayscale images, the program divides the image into 8x8 blocks, applies discrete cosine transform (DCT) to each block, quantizes the DCT coefficients using a quantization matrix, and reconstructs the image. Reducing the quantization factor increases compression but decreases quality. For color images, the program converts RGB to YIQ components, applies separate DCT and quantization to each channel, then reconstructs the RGB image. Running the programs with different quantization factors for Y, I, and Q channels affects the range of quantized DCT coefficients and perceived image quality.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as DOC, PDF, TXT or read online from Scribd
    Download as doc, pdf, or txt
    75 views5 pages

    DSP Lab 13

    Uploaded by

    mpssassygirl
    The document describes MATLAB programs for grayscale and color JPEG image compression. For grayscale images, the program divides the image into 8x8 blocks, applies discrete cosine transform (DCT) to each block, quantizes the DCT coefficients using a quantization matrix, and reconstructs the image. Reducing the quantization factor increases compression but decreases quality. For color images, the program converts RGB to YIQ components, applies separate DCT and quantization to each channel, then reconstructs the RGB image. Running the programs with different quantization factors for Y, I, and Q channels affects the range of quantized DCT coefficients and perceived image quality.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as DOC, PDF, TXT or read online from Scribd
    Download as doc, pdf, or txt
    of 5

    MATLAB

    Lab 13 Quality Effects on JPEG Image Compression Using MATLAB


    OBJECTIVES
    Investigate the grayscale JPEG image compression with the quality effects
    Investigate the color JPEG image compression with the quality effects.
    PROCEDURE:
    Part A: Grayscale JEPG Image
    The grayscale JPEG image compression includes the following steps. a. Dividing the image into 8x8
    blocks. b. Transforming each block using the discrete-cosine transform. c. Scaling and Quantizing DCT
    coefficients. d. Lossless-compressing the quantized DCT coefficients. Note that we do not pursue the loss
    compression here assuming that the lossless compressed DCT coefficients will be exactly recovered during
    the decoding stage. Examine the following sample MATLAB program provided by your instructor.
    close all; clear all; clc
    % This test program is a part of simulation for Grayscale JPEG image
    % comression.
    % image size: 240x320
    X=imread('cruiseog','TIFF');
    %note the image size= 240x320
    Y=rgb2gray(X);
    %Y =grayscale image
    size(Y)
    Q=[ 16 11 10 16 24 40 51 61;
    %quantization quality factor
    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];
    factor=1;
    Q=factor*Q;
    Order=[ 0 1 5 6 14 15 27 28;
    %zig-zag path
    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];
    count=0;
    % number of image blocks; total number of blocks =30x40=1200
    for i=1:30
    for j=1:40
    count=count+1;
    P(1:8,1:8,count)=Y((i-1)*8+1:i*8,(j-1)*8+1:j*8);
    end
    end
    for m=1:1200
    F(:,:,m)=round(dct2(P(:,:,m))); % 2-D DCT transform

    MATLAB

    FF(:,:,m)=round(F(:,:,m) ./Q); % quantization of DCT coefficients


    end
    %Coding and lossless compression. (not included)
    %decoding
    for m=1:1200
    FC(:,:,m)=FF(:,:,m) .*Q;
    % recover 2-D DCT coefficients
    PP(:,:,m)=round(idct2(FC(:,:,m))); % recover the 8x8 image image
    err(:,:,m)=single(PP(:,:,m))-single(P(:,:,m)); % quantization vector
    end
    count=0;
    for i=1:30
    for j=1:40
    count=count+1;
    YY((i-1)*8+1:i*8,(j-1)*8+1:j*8)=PP(1:8,1:8,count); % recover image
    end
    end
    YY=uint8(YY); % level adjustment
    figure
    subplot(1,2,1),imshow(Y); % original image
    title('Orignal image');
    subplot(1,2,2),imshow(YY) % JPEG image
    title('JPEG compressed image');
    In the program, Q contains the scaled factors while FF(:,:m) contains the quantized DCT coefficients. If we
    reduce the range of the FF(:,:,m), we will achieve more compression with trading off the image quality.
    Run the MATLAB program to complete the following table using the specified scale factor, and examine
    the range of the quantized DCT coefficients and image quality.
    factor
    1
    2
    5
    10
    20

    MATLAB>max(FF(:)

    Quality

    Write a brief summary.

    Instructor verification ________________


    Part B: Color JPEG Image
    Examine the following MATLAB program provided by your instructor. The principles of the color JPEG
    image compression include: a. converting RGB image components to YIQ image components. b. Applying
    DCT transform to each the YIQ image components, respectively. c. Scaling and quantizing YIQ image
    components using the corresponding quantization factor (Q for Y image while Qc for I and Q images in the
    program), respectively. d. lossless coding the DCT coefficients (we do not pursue here).

    MATLAB

    close all; clear all; clc


    % This test program is a part of simulation for color JPEG image
    % compression.
    % image size: 240x320x3
    X=imread('grandorg','TIFF');
    X=single(X); % single precision
    Y=rgb2gray(X); % Y -channel
    Q=[ 16 11 10 16 24 40 51 61; % quality factor for Y components
    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];
    QC=[17 18 24 47 99 99 99 99; % quality factor of I or Q components
    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];
    factorY=1;
    Q=factorY*Q;
    factorC=1;
    QC=factorC*QC;
    Order=[ 0 1 5 6 14 15 27 28; %zig-zag path
    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];
    %transform to YIQ
    YIQ(:,:,1)=0.299*X(:,:,1)+0.587*X(:,:,2)+0.114*X(:,:,3);
    YIQ(:,:,2)=0.596*X(:,:,1)-0.274*X(:,:,2)-0.322*X(:,:,3);
    YIQ(:,:,3)=0.212*X(:,:,1)-0.523*X(:,:,2)+0.311*X(:,:,3);
    count=0; % number of 8x8x3 image blocks for Y components
    for i=1:30
    for j=1:40
    count=count+1;
    P(1:8,1:8,count)=YIQ((i-1)*8+1:i*8,(j-1)*8+1:j*8,1);
    end
    end
    for m=1:1200 % operating for 30x40x3 blocks =1200 blocks
    F(:,:,m)=round(dct2(P(:,:,m))); % DCT transform
    FF(:,:,m)=round(F(:,:,m) ./Q); % DCT coefficient quantization
    end
    count=0; % operate for I componenets
    for i=1:30

    MATLAB

    for j=1:40
    count=count+1;
    PI(1:8,1:8,count)=YIQ((i-1)*8+1:i*8,(j-1)*8+1:j*8,2);
    end
    end
    for m=1:1200
    F(:,:,m)=round(dct2(PI(:,:,m))); %DCT transform
    FFI(:,:,m)=round(F(:,:,m)./QC); % DCT coefficient quantization
    end
    count=0;
    % operating for Q component
    for i=1:30
    for j=1:40
    count=count+1;
    PQ(1:8,1:8,count)=YIQ((i-1)*8+1:i*8,(j-1)*8+1:j*8,3);
    end
    end
    for m=1:1200
    F(:,:,m)=round(dct2(PQ(:,:,m))); %DCT transform
    FFQ(:,:,m)=round(F(:,:,m)./QC); % DCT coefficient quantzation
    end
    % Further coding and lossless compression (not included)
    %decoding
    for m=1:1200
    FC(:,:,m)=FF(:,:,m) .*Q;
    % recovering DCT coefficients
    PP(:,:,m)=round(idct2(FC(:,:,m))); % recovering 8x8 image block for Y component
    end
    count=0;
    for i=1:30
    for j=1:40
    count=count+1;
    YY((i-1)*8+1:i*8,(j-1)*8+1:j*8,1)=PP(1:8,1:8,count); % Y components
    end
    end
    for m=1:1200 %recoverinbg the I component
    FIC(:,:,m)=FFI(:,:,m).*QC;
    PPI(:,:,m)=round(idct2(FIC(:,:,m)));
    end
    count=0;
    for i=1:30
    for j=1:40
    count=count+1;
    YY((i-1)*8+1:i*8,(j-1)*8+1:j*8,2)=PPI(1:8,1:8,count);
    end
    end
    for m=1:1200 %recoverinbg the Q component
    FQC(:,:,m)=FFQ(:,:,m).*QC;
    PPQ(:,:,m)=round(idct2(FQC(:,:,m)));
    end

    MATLAB

    count=0;
    for i=1:30
    for j=1:40
    count=count+1;
    YY((i-1)*8+1:i*8,(j-1)*8+1:j*8,3)=PPQ(1:8,1:8,count);
    end
    end
    %transform to RGB
    RGB(:,:,1)=1.0*YY(:,:,1)+0.956*YY(:,:,2)+0.621*YY(:,:,3);
    RGB(:,:,2)=1.0*YY(:,:,1)-0.272*YY(:,:,2)-0.647*YY(:,:,3);
    RGB(:,:,3)=1.0*YY(:,:,1)-1.106*YY(:,:,2)+1.703*YY(:,:,3);
    X=uint8(X); %level adjustment
    RGB=uint8(RGB);%level adjustment
    subplot(1,2,1),imshow(X);title('Original image');
    subplot(1,2,2),imshow(RGB); title('JPEG image');
    Run the MATLAB program to complete the following table using the specified scale factors, and examine
    the ranges of the quantized DCT coefficients and image quality.
    Factor Y
    1
    2
    5
    10
    4

    Factor C
    1
    2
    4
    8
    10

    MAT>max(FF(:) )

    Write a brief summary.


    Instructor verification ________________

    MAT>max(FFI(:))

    MAT>max(FFQ(:))

    Quality

    You might also like