压缩感知与单像素成像ppt课件.pptx

,| Compressive Sensing of Images,关 晨孙 晓 雯李 显 业,1 . Introduction,Signal,?,Signal,Emmanuel Cands,ZhexuanTao,2 . Principle,S(n*1),(n*n),X(n*1)k sparseness,(m*n),Y(m*1),Y=S=X,Where m satisfied the equation m C*k*log(n/k) (C is a constant),Usually, is a known basis, knowledge of X is equivalent to knowledge of S.,How to solve X from the function Y=X. That is a convex optimization problem,2 . Principle,Generally compressive sensing relied on the assumption that the solution to the l1 minimization problem provides the correct solution and is computationally feasible. However work has been done to find alternative algorithms that are faster or give superior reconstruction performance.,Matching Pursuit (MP),Orthogonal Matching Pursuit (OMP),Stagewise Orthogonal Matching Pursuit (StOMP),Compressive Sampling Matching Pursuit (CoSaMP),Two-Step Iterative Shrinkage Thresholding (TwIST),2 . Principle,Initial image,sparsetransform,FFT DCTDWT ,Sparse image,Measurement matrix,compressed image,Measurement matrix,Sparse image,Inverse transformation,Reconstruction algorithms,3 . Simulation,Initial image,DWT,Sparse,Measurement,Reconstruction,DCT,Reconstruction,Sparse,4 . Application,Single pixel camera (computational ghost imaging),Y=X,4 . Application,Single pixel camera (Compressive ghost imaging),Y=S=X,4 . Application,Initial image,Reconstruction,filtering,compressionratio: 7.93%,PSNR： 62.048 dB,4 . Application,4 . Application,Thanks,