README.md

Enhanced-3DTV

The motivation map of Enhanced-3DTV is as follows.

The matlab code of paper ''Enhanced 3DTV Regularization and Its Applications on Hyper-spectral Image Denoising and Compressed Sensing''

There are Two tasks: Compress sensing & Denoising

structure

Compress sensing

The structure of Compress sensing is:

• compared method
  • JtenRe3DTV
  • KSC
  • LR&TV
  • SLNTCS
  • SpaRCS
• Enhanced3DTV in the paper
  • demo_EnhancedTV_CS.m
  • EnhancedTV_CS.m
  • quality assess
• quality assess
• demo.m

# run "demo.m" in command line window of matlab to test all codes
$ demo.m
# run the code inside "Enhanced3DTV in the paper" to see the performances of Enhanced 3DTV in compress sensing tasks.
$ demo_EnhancedTV_CS.m

denoising

The structure of Denoising is:

• compete code
  • ALM_RPCA
  • BM4D
  • LLRT
  • LRTDTV
  • LRTV
  • TDL
  • WNNM
  • WSNM_RPCA
• Enhanced3DTV in the paper
  • EnhancedTV.m
  • TV_operator
  • quality assess
  • simulation_case1_demo.m
  • simulation_case2_demo.m
  • simulation_case3_demo.m
  • simulation_case4_demo.m
  • simulation_case5_demo.m
  • simulation_case6_demo.m
• quality assess
• Demo_simulation_case1.m
• Demo_simulation_case3.m
• RunAllMethod.m

API of all methods are list in "RunAllMethod.m"

# run "Demo_simulation_case1.m" and "Demo_simulation_case3.m"to test all the code. For example,
$ Demo_simulation_case3.m
# run the code inside "Enhanced3DTV in the paper" to see the performances of Enhanced 3DTV in Denoise tasks. For example,
$ simulation_case3_demo.m

supplemental.pdf

Here, we show some visual restoration of all methods The denoising performance of all methods on IndianPines simulation data. The compress sensing restore performance of all methods on dcmall data. The compress sensing restore performance of all methods on lowal altitude data.

More experiment results and the proof of the equivalence are list in supplemental.pdf