CSLDS_Demo.rar_DEMO_动态压缩感知

******* Run demo.m It loads a dynamic texture of a candle and attempts to reconstruct it from noiselet measurements. - Read paper at http://www.ece.rice.edu/~as48/research/cslds for additional details. - You might need spgl1 package if you want to use the group-sparsity version of CSLDS. Contents: 1) "cosamp" folder contains two codes for model-based CS. - Cosamp_mean_cslds.m is the version for the mean+LDS model. This is the preferred implementation - Cosamp_groupsparsity.m is the version with a group-sparsity term. The command spg_group in the toolbox SPGL1, in most cases, works better than this. - Both codes can handle functional handles for the measurement operator. 2) run_cslds.m is the main workhorse script. demo.m gives an idea of how to use it. 3) "functions" folder - Mainly a list of measurement and adjoint operators specific to the CSLDS framework. A lot different sparsifying basis are defined here. 4) "utility" folder - Implementation of noiselets. Credits to Dr. Jason Laska (laska@rice.edu) 5) "dyntex" has frames of a dynamic texture from the DynTex database