matlab-基于haar和小波变换的一维ECG信号压缩以及二维图像压缩matlab仿真,对比不同压缩率下的数据图像质量-源码

image = imread('cheetah.png'); grayimage = rgb2gray(image); imshow(grayimage) title('compression ratio = 0%') [a,h,v,d] = haart2(grayimage); %% Plot Different Compression Ratios % [threshold5, data_sort5] = FindThreshold2D(a,h,v,d,0.5); % [rec5,a_rec5,h_rec5,v_rec5,d_rec5] = reconstruction2D(a,h,v,d,threshold5); % figure % subplot(1,2,1) % imshow(uint8(rec5)) % title('compression ratio = 50%') % [threshold8, data_sort8] = FindThreshold2D(a,h,v,d,0.8); % [rec8,a_rec8,h_rec8,v_rec8,d_rec8] = reconstruction2D(a,h,v,d,threshold8); % subplot(1,2,2) % imshow(uint8(rec8)) % title('compression ratio = 80%') % % figure % [threshold9, data_sort9] = FindThreshold2D(a,h,v,d,0.9); % [rec9,a_rec9,h_rec9,v_rec9,d_rec9] = reconstruction2D(a,h,v,d,threshold9); % subplot(1,2,1) % imshow(uint8(rec9)) % title('compression ratio = 90%') % [threshold98, data_sort98] = FindThreshold2D(a,h,v,d,0.98); % [rec98,a_rec98,h_rec98,v_rec98,d_rec98] = reconstruction2D(a,h,v,d,threshold98); % subplot(1,2,2) % imshow(uint8(rec98)) % title('compression ratio = 98%') % % figure % [threshold99, data_sort99] = FindThreshold2D(a,h,v,d,0.99); % [rec99,a_rec99,h_rec99,v_rec99,d_rec99] = reconstruction2D(a,h,v,d,threshold99); % subplot(1,2,1) % imshow(uint8(rec99)) % title('compression ratio = 99%') % [threshold995, data_sort995] = FindThreshold2D(a,h,v,d,0.995); % [rec995,a_rec995,h_rec995,v_rec995,d_rec995] = reconstruction2D(a,h,v,d,threshold995); % subplot(1,2,2) % imshow(uint8(rec995)) % title('compression ratio = 99.5%') % % figure % [threshold999, data_sort999] = FindThreshold2D(a,h,v,d,0.999); % [rec999,a_rec999,h_rec999,v_rec999,d_rec999] = reconstruction2D(a,h,v,d,threshold999); % subplot(1,2,1) % imshow(uint8(rec999)) % title('compression ratio = 99.9%') % [threshold9995, data_sort9995] = FindThreshold2D(a,h,v,d,0.9995); % [rec9995,a_rec9995,h_rec9995,v_rec9995,d_rec9995] = reconstruction2D(a,h,v,d,threshold9995); % subplot(1,2,2) % imshow(uint8(rec9995)) % title('compression ratio = 99.95%') % % figure % [threshold9997, data_sort9997] = FindThreshold2D(a,h,v,d,0.9997); % [rec9997,a_rec9997,h_rec9997,v_rec9997,d_rec9997] = reconstruction2D(a,h,v,d,threshold9997); % subplot(1,2,1) % imshow(uint8(rec9997)) % title('compression ratio = 99.97%') % [threshold9999, data_sort9999] = FindThreshold2D(a,h,v,d,0.9999); % [rec9999,a_rec9999,h_rec9999,v_rec9999,d_rec9999] = reconstruction2D(a,h,v,d,threshold9999); % subplot(1,2,2) % imshow(uint8(rec9999)) % title('compression ratio = 99.99%') %% Compute PSNR, MSE and SSIM ratio = (0.01:0.005:1)'; ratio = [ratio; (0.9950:0.0001:0.9999)']; [psnr,mse,ssim] = ComputePSNRs2D(double(grayimage),a,h,v,d,ratio); figure plot(ratio,psnr,'x') xlabel('compression ratio') ylabel('PSNR') figure plot(ratio,mse,'x') xlabel('compression ratio') ylabel('MSE') figure plot(ratio,ssim,'x') xlabel('compression ratio') ylabel('SSIM') % Function to reconstruct a 2D image given a specific threshold % INPUTS % a,h,v,d: detail and approximation coefficient % threshold: coefficients smaller than threshold will be set to zero % OUTPUTS % rec: reconstructed signal in 2D % a,h,v,d: updated detail and approximation coefficient function [rec,a,h,v,d] = reconstruction2D(a,h,v,d,threshold) count = 0; total = 0; % total nonzero elements if a ~= 0 if abs(a) < threshold a = 0; count = count + 1; end total = total + 1; end for i = 1:length(d) for j = 1:(size(d{i},1)*size(d{i},2)) if d{i}(j) ~= 0 if abs(d{i}(j)) < threshold d{i}(j) = 0; count = count + 1; end total = total +1; end if h{i}(j) ~= 0 if abs(h{i}(j)) < threshold h{i}(j) = 0; count = count + 1; end total = total + 1; end if v{i}(j) ~= 0 if abs(v{i}(j)) < threshold v{i}(j) = 0; count = count + 1; end total = total + 1; end end end rec = ihaart2(a,h,v,d); compression_rate = count/total end % Function to find a desired threshold given a specific compression ratio % INPUTS % a,h,v,d: detail and approximation coefficient % ratio: desired compression ratio % OUTPUTS % threshold: desired threshold value % data_sort: 1D array of sorted coefficients function [threshold,data_sort] = FindThreshold2D(a,h,v,d,ratio) n = size(h{1},1)*2; data = zeros(n,n); data(1) = a; index = 2; for i=1:length(d) for j = 1:(size(d{i},1)*size(d{i},2)) data(index:index+2) = [h{i}(j), v{i}(j),d{i}(j)]; index = index + 3; end end data_sort = sort(abs(data(:))); start_index = sum(data_sort==0); threshold = data_sort(start_index + ceil(ratio*(length(data_sort)-start_index))); end % Function to compute PSNR, MSE and SSIM given a range of ratios % INPUTS % data: a 1D input signal % ratio: a 1D array of compression ratio % a: approximation coefficient % D: object that stores details coefficient at all levels % OUTPUTS % PSNRs: the corresponding PSNR values % MSEs: the corresponding PSNR values % SSIMs: the corresponding PSNR values function [PSNRs,MSEs,SSIMs] = ComputePSNRs2D(data,a,h,v,d,ratio) PSNRs = zeros(length(ratio),1); MSEs = zeros(length(ratio),1); SSIMs = zeros(length(ratio),1); for i = 1:length(ratio) [threshold, ~] = FindThreshold2D(a,h,v,d,ratio(i)); rec = reconstruction2D(a,h,v,d,threshold); MSE = sum(sum((rec - data).^2))./(size(data,1)*size(data,2)); PSNR = 10*log10(1./MSE); SSIM = ssim(rec,data); MSEs(i) = MSE; PSNRs(i) = PSNR; SSIMs(i) = SSIM; end end