Switching Bilateral Filter Noise Removal matlab code

Here we discuss about Switching Bilateral Filter Noise Removal step by step.

1.Read the image pixel by pixel (i,j).
2. Each pixel we construct 4 sub windows.

For a (2N+1) *(2N+1) window we divide the window into four (N+1)*(N+1) subwindows .


case N = 2

3.Find the Sorted Quadrant Median Vector (SQMV) for all subwindows.

SQM1, SQM2, SQM3 and SQM4 are the medians 

4. Then we find the regions from above all SQM.

Uniform Region

Diagonal edge in dark side

Diagonal edge in dark side

Veritical edge

Horizontal edge

Diagonal line

Gradual chage edge 

Texture

5. Then we find the reference median (SQMR)  based on above regions.

6. Then apply Bilateral Filter.

7. The apply noise detection .

Code:

clc

clear all

close all

close all hidden

warning off

con = 1; % con1 for 

img = imread('lena.jpg');

figure;imshow(img);

title('Input Image');

img = imnoise(img,'salt & pepper',0.01); % salt and pepper

Tk1 = 30; % for salt and paper noise..

Tk2 = 15; % for salt and paper noise..

% img = imnoise(img,'gaussian',0,0.01); % gaussian noise

% Tk1 = 25; % for Gaussian noise..

% Tk2 = 5; % for Gaussian noise..

img = double(img);

figure;imshow(uint8(img));

title('Noisy Image');

ext_filt = img;

sigma_R = 40;

[r c] = size(img);

N = 2;

pi_val = 25;

h = waitbar(0,'Applying Switching bilateral filter..');

L = 1;

for i = (N+1):r-(N+1)

    waitbar(L/length((N+1):r-(N+1)));

    L = L+1;

    for j = (N+1):c-(N+1)

        [m0,m1,m2,m3,m4] = meadian_calc(i,j,img,N);

        dav = dav_valc(m1,m2,m3,m4,img,i,j);

        SQMDB = m4-m1;

        SQMDC = m3-m2;

        SQMR = SQMR_calc(SQMDC,m1,m2,m3,m4,dav,pi_val);

        [s1 s2] = find_S1_S2(img,i,j,SQMR,Tk1,Tk2);

        if(dav~=0)

            sigma_S = 3;

        else

            sigma_S = 1;

        end

        f_x_y = f_of_x_y(i,j,N,img,sigma_S,sigma_R,s2,SQMR);

        ext_filt(i,j) = f_x_y;

    end

end

close(h); 

figure;

imshow(uint8(ext_filt));%output...

Referencce : Switching Bilateral Filter With a Texture/Noise Detector for Universal Noise Removal Chih-Hsing Lin, Jia-Shiuan Tsai, and Ching-Te Chiu Transactions on: Image Processing, IEEE Journals 2010