ACS_Net_MultiScale.m

% This is an implement of strategy 2(add Sample strategy );
clear
t1 = clock;
delete('SampleRate.txt');

%this global param controls the block image order;
saveResult = 1;
showResult = 1;
global useGPU;
useGPU = 0;
global blockNo; 
blockNo = 0;
%blocksize of blockproc;
blockSize = 64;
%%
imageNo = 450;
%Choose image;
image = imread(['../BigData/datasets/reference-890/GroundTruth/','image',num2str(imageNo),'.png']);
if size(image,3)==3
    image = rgb2ycbcr(image);
    image = im2double(image(:, :, 1));
else
    image = im2double(image);
end
% convert image -> can be mod by 32;
image =modcrop(image,32); 

%%
%base Sampling by SR of 0.05;
%基础采样部分：
baseInitRec = ACS_Net_initRec(image,0.05); 

%%
%compute Entropy matrix by baseInitRec！！
%根据初始重构分块计算图像熵：
%注意：使用"entropy"时,使用entropy函数来计算熵; 使用"entropyfilt"时,使用entropyfilt函数来计算熵; 两者结果可能不同；
fun = @(block_struct)ComputeEntropy(block_struct.data,"entropy");
entropyMatrix = blockproc(baseInitRec,[blockSize,blockSize],fun);
%convert the entropyMatrix to a 1D entropyVector;
entropyVector = getEntropyVector(entropyMatrix);
%sort the entropyVector;
sortedEntropyVector = sort(entropyVector);

%%
%补充采样部分：
%分块分不同采样率补充采样,获得残差；这里需要用到entropyVector和sortedEntropyVector;
fun = @(block_struct)MutiScaleResidual(block_struct.data,entropyVector,sortedEntropyVector); 
initEnResidual  = blockproc(image,[blockSize blockSize],fun);
%基准初始重构+补充采样的残差=恢复的初始重构
TotalinitRec = initEnResidual + baseInitRec; 

%%
%deep Reconstruction!
%深度重构：
deepReconstruction = initial2Deep(TotalinitRec);

%%
%evaluate the quality of the result  && show result
if showResult
    [psnr_base,ssim_base]             = Cal_PSNRSSIM(im2uint8(image) ,im2uint8(baseInitRec),0,0);
    [psnr_totalInit,ssim_totalInit]  = Cal_PSNRSSIM(im2uint8(image) ,im2uint8(TotalinitRec),0,0);
    [psnr_deep,ssim_deep]            = Cal_PSNRSSIM(im2uint8(image) ,im2uint8(deepReconstruction),0,0);
    
    subplot(1,4,1);
    imshow(baseInitRec);
    title('0.05采样率的基准初始重构');
    subplot(1,4,2);
    imshow(TotalinitRec);
    title('补充采样后的初始重构');
    subplot(1,4,3);
    imshow(deepReconstruction);
    title('深度重构');
    subplot(1,4,4);
    imshow(image);
    title('原图');
    
    %display the average Sample Rate;
    filename = 'SampleRate.txt';
    SRs=textread(filename,'%f');
    AvgSampleRate = mean(SRs);
    fprintf('Average Sample Rate = %g\n',AvgSampleRate);
end

t2 = clock;
runtime = etime(t2,t1);
%%
%save Initial Result && Deep Result!
if saveResult == 1
    imwrite(TotalinitRec,              ['../BigData/datasets/reference-890/InitialResult/','image',num2str(imageNo),'.png']); 
    imwrite(deepReconstruction,['../BigData/datasets/reference-890/DeepResult/','image',num2str(imageNo),'.png']); 
end