how can I implement image fusion algorithm by using curvelet transform?

Image fusion is the process of combining two images to form only one image which should contain information of both the images.

Answer 1

[surya mahajan]

clearall;

closeall;

%%

[filename filepath]=uigetfile('*.jpg','select the input chromatic image');

filedir1=strcat(filepath,'\',filename);

[filename filepath]=uigetfile('*.jpg','select the input spectral image');

filedir2=strcat(filepath,'\',filename);

im1=imresize(imread(filedir1),[512,512]);

im2=imresize(imread(filedir1),[512,512]);

figure,imshow(im2);

title('inputimage 1-chromatic');

figure,imshow(im2);

tittle('input image 2-multi spectral');

no_of_level=3;

curve_wavelet_2d=CURVELET_PROCESS(im1,1,no_of_level);

for k3=1:3

curve_wavelet_3d=CURVELET_PROCESS(im2(:,:,k3),1,no_of_level);

end

%%

fork5=1:3

curv_coeff=curve_wavelet_3d{k5};

fork6=1:lenght(curv_coeff)-2

tmp_coeff=curv_coeff{k6};

tmp_2dcoeff=curve_wavelet_2d{k6};

for k7=1:length(tmp_coeff)

coeff3d=temp_coeff{k7};

coeff2d=tmp_2dcoeff{k7};

locl=find(coeff2d<1);

coeff3d(locl)=coeff2d(locl);

res..coeff3{k7}=coeff3d;

end

res..coeff2{k6}=res_coeff3;

res_coeff3=[];

end

res_coeff2{k6+1}=curv_coeff{k6+1};

res_coeff2{k6+2}=curv_coeff{k6+2};

final..res..coeff{k5}=res_coeff2;

end

fork10=1:3

imgx=INVERSE_CURVELET_PROCESS(FINAL_RES_COEFF{K10});

imgout(:,:,k10)=imgx;

end

figure,imshow(uint8(imgout));

tittle('resultimage');

%%

resout=uint8(imgout);

for km=1:3

disp(strcat('band--',num2str(km)))

mseerroe=mean(mean(im2(:,:,km)-uint8(imgout(:,:,km))).^2));

MSE_VAL=mseerroe

PSNR_VAL=20*log10(255/mseerroe)

ENTROPY_VAL=entropy_func(double(resout(:,:,km)))

f1img=im2(:,:,km);

f2img=resout(:,:,km);

corrval=xcorr(f1img(:),f2img(:));

CROSS_CORRELATION=max(corrval)

[fx,fy]=gradient(double(resout(:,:,km)));

AVERAGE_GRADIENT_VAL=mean(mean(mean(sqrt((fx.^2+fy.^2)/2))))

UIQI=UniversalImageQualityIndex(double(im2(:,:,km)),double(resout(:,:,km)))

end

[Bhakti Raul]

could find CURVELET_PROCESS and INVERSE_CURVELET_PROCESS functions anywhere

[jyoti khandelwal]

defination of curvelet and inverse curvelet fuction is missing