Scribd
Upload
50 views28 pages

JPEG - Still Image Compression

This document describes the JPEG image compression algorithm. It involves several key steps: (1) converting the image to grayscale and dividing it into 8x8 pixel blocks, (2) performing discrete cosine transform (DCT) and quantization on each block, (3) separating AC and DC components and zigzag scanning, (4) encoding the data using run length and differential coding, (5) decoding, reconstructing the image by inverse DCT, and (6) comparing the original and reconstructed images. The algorithm aims to reduce file size with minimal loss of visual quality for still images.

Uploaded by

Zeeshan Akhtar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
50 views28 pages

JPEG - Still Image Compression

This document describes the JPEG image compression algorithm. It involves several key steps: (1) converting the image to grayscale and dividing it into 8x8 pixel blocks, (2) performing discrete cosine transform (DCT) and quantization on each block, (3) separating AC and DC components and zigzag scanning, (4) encoding the data using run length and differential coding, (5) decoding, reconstructing the image by inverse DCT, and (6) comparing the original and reconstructed images. The algorithm aims to reduce file size with minimal loss of visual quality for still images.

Uploaded by

Zeeshan Akhtar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 28

JPEG Still Image

Compression
Zeeshan Akhtar
Mtech (1st year )
14-LECM 018
GI-4784
A.M.U. aligarh

Encoding

JPEG:Basic Algorithm

Step 1-

Read the image


Convert it to grayscale
Offset level
Determine the size of
image

image = imread('leena.jpg');
gray = rgb2gray(image);
d = double(gray);
D = d-128;%offset
[M,N] = size(D);

Step 2

Zero Padding
Determine the new size
of image

Rem_M = mod(M,8);
Rem_N = mod(N,8);
Extra_M = 8 - Rem_M;
Extra_N = 8 - Rem_N;
if Extra_M==0
M_new = M;
else
M_new = M + Extra_M;
end
if Extra_N==0
N_new = N;
else N_new = N + Extra_N;
end
D_new = padarray(D,[Extra_M
Extra_N],'post');

Step 3
Define Quantization table

Q = [ 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 36 55 64 81 104 113 92
49 64 78 87 103 121 120 101
72 92 95 98 112 100 103 99];

Step 4
Break image in small blocks of 8*8
dimension
Perform DCT on each block
Perform quantization
Separate AC and DC part of the image
Zig zag Scanning

X = M_new/8;
Y = N_new/8;
DC_array = zeros(1,X*Y);
AC_matrix= zeros(X*Y,63);
c=1; d=8; t = 1;
for j = 1:X
a=1;
b=8;
for i = 1:Y
small_D = D_new(c:d,a:b);
dct = round(dct2(small_D));
Quant_D = round(dct./Q);
DC_array(t) = Quant_D(1,1);
AC_matrix(t,1:63) =
zigzag2(Quant_D);
t = t + 1;
a = a+8;
b = b+8;
end
c = c+8;
d = d+8;
end

function file=zigzag(T);
file=[
T(1,2) T(2,1) T(3,1) T(2,2) ...
T(1,3) T(1,4) T(2,3) T(3,2) T(4,1) ...
T(5,1) T(4,2) T(3,3) T(2,4) T(1,5) ...
T(1,6) T(2,5) T(3,4) T(4,3) T(5,2) ...
T(6,1) T(7,1) T(6,2) T(5,3) T(4,4) ...
T(3,5) T(2,6) T(1,7) T(1,8) T(2,7) ...
T(3,6) T(4,5) T(5,4) T(6,3) T(7,2) ...
T(8,1) T(8,2) T(7,3) T(6,4) T(5,5) ...
T(4,6) T(3,7) T(2,8) T(3,8) T(4,7) ...
T(5,6) T(6,5) T(7,4) T(8,3) T(8,4) ...
T(7,5) T(6,6) T(5,7) T(4,8) T(5,8) ...
T(6,7) T(7,6) T(8,5) T(8,6) T(7,7) ...
T(6,8) T(7,8) T(8,7) T(8,8)]

Step5

Perform Differential Coding


on DC part separated

coded_DC = zeros(1,X*Y);
L1 = (X*Y)-1;
for i=1:L1
coded_DC(i+1) =
DC_array(i+1)- DC_array(i);
end
coded_DC(1) = DC_array(1);

Step 6

Run length coding of AC


separated

L2 = (X*Y);
for i=1:L2
AC_array = AC_matrix(i,:);
coded_AC(i,:) = rle(AC_array);
end

Decoding

Step 1- AC decoding(Run length)

for i = 1:L2
decoded_AC(i,:) =
rle(coded_AC(i,:));
end

Step2- Diffrential Decoding

decoded_DC(1) = coded_DC(1);
for i=1:L1
decoded_DC(i+1) = decoded_DC(i)
+ coded_DC(i+1);
end

Step3

IDCT
INVERSE ZIG ZAG SCANNING
IMAGE RECONSTRUCTION

lowlevel_image = zeros(N_new,M_new);
a=1;
b=8;
c=1;
d=8;
t=1;
for i = 1:X
for j= 1:Y
AC_DC_array = AC_DC(t,:);
q=zeros(8,8);
q = invzigzag(AC_DC_array);
Inv_Q = round(q.*Q);
lowlevel_image(a:b,c:d) = round(idct2(Inv_Q));
c=c+8;
d=d+8;
t=t+1;
end
c=1;
d=8;
a=a+8;
b=b+8;
end

Inverse ZiG- ZaG Mapping


function inv=invzigzag(T);
inv=[
T(1,1)
T(1,16) T(1,28)
T(1,3)
T(1,27) T(1,30)
T(1,4)
T(1,31) T(1,42)
T(1,10)
T(1,41) T(1,45)
T(1,11)
T(1,46) T(1,53)
T(1,21)
T(1,52) T(1,56)
T(1,22)
T(1,57) T(1,60)
T(1,36)
T(1,59) T(1,63)

T(1,2)
T(1,29)
T(1,5)
T(1,43)
T(1,9)
T(1,44)
T(1,12)
T(1,54)
T(1,20)
T(1,55)
T(1,23)
T(1,61)
T(1,35)
T(1,62)
T(1,37)
T(1,64)

T(1,6)

T(1,7)

T(1,15)

T(1,8)

T(1,14) T(1,17)

T(1,13) T(1,18) T(1,26)


T(1,19) T(1,25) T(1,32)
T(1,24) T(1,33) T(1,40)
T(1,34) T(1,39) T(1,47)
T(1,38) T(1,48) T(1,51)
T(1,49) T(1,50) T(1,58)
]

Step 4 - Inverse Offset

uplevel_image = lowlevel_image + 128;


Image = uint8(uplevel_image);

Step 5
Display both the original and
compressed images
Compare both the images
Find and display the difference between
two.

figure;
imshow(gray);
title('original image')
figure;
imshow(Image);
title('after decoding')
difference= (d)(uplevel_image);
figure;
imshow(difference);

Thank You

You might also like