主要內容:算法
分段弱正交匹配追蹤(Stagewise Weak OMP)能夠說是StOMP的一種修改算法,它們的惟一不一樣是選擇原子時的門限設置,這能夠下降對測量矩陣的要求。咱們稱這裏的原子選擇方式爲"弱選擇"(Weak Selection),StOMP的門限設置由殘差決定,這對測量矩陣(原子選擇)提出了要求,而SWOMP的門限設置則對測量矩陣要求較低(原子選擇相對簡單、粗糙)。性能
SWOMP的算法流程:測試
function [ theta ] = CS_SWOMP( y,A,S,alpha ) % CS_SWOMP % Detailed explanation goes here % y = Phi * x % x = Psi * theta % y = Phi*Psi * theta % 令 A = Phi*Psi, 則y=A*theta % S is the maximum number of SWOMP iterations to perform % alpha is the threshold parameter % 如今已知y和A,求theta % Reference:Thomas Blumensath,Mike E. Davies.Stagewise weak gradient % pursuits[J].IEEE Transactions on Signal Processing,2009,57(11):4333-4346. if nargin < 4 alpha = 0.5; %alpha範圍(0,1),默認值爲0.5 end if nargin < 3 S = 10; %S默認值爲10 end [y_rows,y_columns] = size(y); if y_rows<y_columns y = y'; %y should be a column vector end [M,N] = size(A); %傳感矩陣A爲M*N矩陣 theta = zeros(N,1); %用來存儲恢復的theta(列向量) Pos_theta = []; %用來迭代過程當中存儲A被選擇的列序號 r_n = y; %初始化殘差(residual)爲y for ss=1:S %最多迭代S次 product = A'*r_n; %傳感矩陣A各列與殘差的內積 sigma = max(abs(product)); Js = find(abs(product)>=alpha*sigma); %選出大於閾值的列 Is = union(Pos_theta,Js); %Pos_theta與Js並集 if length(Pos_theta) == length(Is) if ss==1 theta_ls = 0; %防止第1次就跳出致使theta_ls無定義 end break; %若是沒有新的列被選中則跳出循環 end %At的行數要大於列數,此爲最小二乘的基礎(列線性無關) if length(Is)<=M Pos_theta = Is; %更新列序號集合 At = A(:,Pos_theta); %將A的這幾列組成矩陣At else%At的列數大於行數,列必爲線性相關的,At'*At將不可逆 if ss==1 theta_ls = 0; %防止第1次就跳出致使theta_ls無定義 end break; %跳出for循環 end %y=At*theta,如下求theta的最小二乘解(Least Square) theta_ls = (At'*At)^(-1)*At'*y; %最小二乘解 %At*theta_ls是y在At列空間上的正交投影 r_n = y - At*theta_ls; %更新殘差 if norm(r_n)<1e-6 %Repeat the steps until r=0 break; %跳出for循環 end end theta(Pos_theta)=theta_ls;%恢復出的theta end
%壓縮感知重構算法測試 clear all;close all;clc; M = 128; %觀測值個數 N = 256; %信號x的長度 K = 30; %信號x的稀疏度 Index_K = randperm(N); x = zeros(N,1); x(Index_K(1:K)) = 5*randn(K,1); %x爲K稀疏的,且位置是隨機的 Psi = eye(N); %x自己是稀疏的,定義稀疏矩陣爲單位陣x=Psi*theta Phi = randn(M,N)/sqrt(M); %測量矩陣爲高斯矩陣 A = Phi * Psi; %傳感矩陣 y = Phi * x; %獲得觀測向量y %% 恢復重構信號x tic theta = CS_SWOMP( y,A); x_r = Psi * theta; % x=Psi * theta toc %% 繪圖 figure; plot(x_r,'k.-'); %繪出x的恢復信號 hold on; plot(x,'r'); %繪出原信號x hold off; legend('Recovery','Original') fprintf('\n恢復殘差:'); norm(x_r-x) %恢復殘差
一、門限參數a分別爲0.1-1.0時,不一樣稀疏信號下,測量值M與重構成功機率的關係:ui
clear all;close all;clc; %% 參數配置初始化 CNT = 1000; %對於每組(K,M,N),重複迭代次數 N = 256; %信號x的長度 Psi = eye(N); %x自己是稀疏的,定義稀疏矩陣爲單位陣x=Psi*theta alpha_set = 0.1:0.1:1; K_set = [4,12,20,28,36]; %信號x的稀疏度集合 Percentage = zeros(N,length(K_set),length(alpha_set)); %存儲恢復成功機率 %% 主循環,遍歷每組(alpha,K,M,N) tic for tt = 1:length(alpha_set) alpha = alpha_set(tt); for kk = 1:length(K_set) K = K_set(kk); %本次稀疏度 %M不必所有遍歷,每隔5測試一個就能夠了 M_set=2*K:5:N; PercentageK = zeros(1,length(M_set)); %存儲此稀疏度K下不一樣M的恢復成功機率 for mm = 1:length(M_set) M = M_set(mm); %本次觀測值個數 fprintf('alpha=%f,K=%d,M=%d\n',alpha,K,M); P = 0; for cnt = 1:CNT %每一個觀測值個數均運行CNT次 Index_K = randperm(N); x = zeros(N,1); x(Index_K(1:K)) = 5*randn(K,1); %x爲K稀疏的,且位置是隨機的 Phi = randn(M,N)/sqrt(M); %測量矩陣爲高斯矩陣 A = Phi * Psi; %傳感矩陣 y = Phi * x; %獲得觀測向量y theta = CS_SWOMP(y,A,10,alpha); %恢復重構信號theta x_r = Psi * theta; % x=Psi * theta if norm(x_r-x)<1e-6 %若是殘差小於1e-6則認爲恢復成功 P = P + 1; end end PercentageK(mm) = P/CNT*100; %計算恢復機率 end Percentage(1:length(M_set),kk,tt) = PercentageK; end end toc save SWOMPMtoPercentage1000 %運行一次不容易,把變量所有存儲下來 %% 繪圖 for tt = 1:length(alpha_set) S = ['-ks';'-ko';'-kd';'-kv';'-k*']; figure; for kk = 1:length(K_set) K = K_set(kk); M_set=2*K:5:N; L_Mset = length(M_set); plot(M_set,Percentage(1:L_Mset,kk,tt),S(kk,:));%繪出x的恢復信號 hold on; end hold off; xlim([0 256]); legend('K=4','K=12','K=20','K=28','K=36'); xlabel('Number of measurements(M)'); ylabel('Percentage recovered'); title(['Percentage of input signals recovered correctly(N=256,alpha=',... num2str(alpha_set(tt)),')(Gaussian)']); end for kk = 1:length(K_set) K = K_set(kk); M_set=2*K:5:N; L_Mset = length(M_set); S = ['-ks';'-ko';'-kd';'-k*';'-k+';'-kx';'-kv';'-k^';'-k<';'-k>']; figure; for tt = 1:length(alpha_set) plot(M_set,Percentage(1:L_Mset,kk,tt),S(tt,:));%繪出x的恢復信號 hold on; end hold off; xlim([0 256]); legend('alpha=0.1','alpha=0.2','alpha=0.3','alpha=0.4','alpha=0.5',... 'alpha=0.6','alpha=0.7','alpha=0.8','alpha=0.9','alpha=1.0'); xlabel('Number of measurements(M)'); ylabel('Percentage recovered'); title(['Percentage of input signals recovered correctly(N=256,K=',... num2str(K),')(Gaussian)']); end
二、稀疏度爲4,12,20,28,36時,不一樣門限參數a下,測量值M與重構成功機率的關係:spa
clear all;close all;clc; load StOMPMtoPercentage1000; PercentageStOMP = Percentage; S = ['-ks';'-ko';'-kd';'-kv';'-k*']; figure; for kk = 1:length(K_set) K = K_set(kk); M_set=2*K:5:N; L_Mset = length(M_set); %ts_set = 2:0.2:3;第3個爲2.4 plot(M_set,Percentage(1:L_Mset,kk,3),S(kk,:));%繪出x的恢復信號 hold on; end load SWOMPMtoPercentage1000; PercentageSWOMP = Percentage; S = ['-rs';'-ro';'-rd';'-rv';'-r*']; for kk = 1:length(K_set) K = K_set(kk); M_set=2*K:5:N; L_Mset = length(M_set); %alpha_set = 0.1:0.1:1;第6個爲0.6 plot(M_set,Percentage(1:L_Mset,kk,6),S(kk,:));%繪出x的恢復信號 hold on; end hold off; xlim([0 256]); legend('StK=4','StK=12','StK=20','StK=28','StK=36',... 'SWK=4','SWK=12','SWK=20','SWK=28','SWK=36'); xlabel('Number of measurements(M)'); ylabel('Percentage recovered'); title(['Percentage of input signals recovered correctly(N=256,ts=2.4,\alpha=0.5)(Gaussian)']);
結論:.net
經過對比能夠看出,整體上講a=0.6時效果較好。3d
對比StOMP中ts=2.4與SWOMP中α=0.6的狀況:StOMP要略好於SWOMP。code