Differences
This shows you the differences between two versions of the page.
Both sides previous revision Previous revision Next revision | Previous revision | ||
teaching:ft1920:praktikum:cs [2020/01/22 23:00] ipa [Projects] |
teaching:ft1920:praktikum:cs [2021/03/02 13:28] (current) |
||
---|---|---|---|
Line 22: | Line 22: | ||
=== Faster FISTA for Wavelet Deblurring (Taken!) === | === Faster FISTA for Wavelet Deblurring (Taken!) === | ||
- | The task ist to implement a fast version of FISTA {{ :teaching:ft1920:vl:cs:files:fasterfista.pdf | Liang, Schönlieb 2019}} and to compare results with the classical version of {{ :teaching:ft1920:vl:cs:files:fista.pdf |FISTA}}. | + | The task is to implement a fast version of FISTA {{ :teaching:ft1920:vl:cs:files:fasterfista.pdf | Liang, Schönlieb 2019}} and to compare results with the classical version of {{ :teaching:ft1920:vl:cs:files:fista.pdf |FISTA}}. |
The regulizer should be chosen as the l1-norm of the {{ :teaching:ft1920:vl:cs:files:wavelet.m.zip | wavelet}} transformed signal. The linear operator should be given as a blurring operator, see below. Try several blurring masks! | The regulizer should be chosen as the l1-norm of the {{ :teaching:ft1920:vl:cs:files:wavelet.m.zip | wavelet}} transformed signal. The linear operator should be given as a blurring operator, see below. Try several blurring masks! | ||
=== FISTA versus the Chambolle-Pock Algorithm for Face Recognition (Taken!) === | === FISTA versus the Chambolle-Pock Algorithm for Face Recognition (Taken!) === | ||
- | The task ist to compare the performance of {{ :teaching:ft1920:vl:cs:files:fista.pdf |FISTA}} | + | The task is to compare the performance of {{ :teaching:ft1920:vl:cs:files:fista.pdf |FISTA}} |
to the {{ :teaching:ft1920:vl:cs:files:chambollepock.pdf |Chambolle-Pock}} algorithm on face recognition. | to the {{ :teaching:ft1920:vl:cs:files:chambollepock.pdf |Chambolle-Pock}} algorithm on face recognition. | ||
To apply FISTA you need to consider the Lagrangian formulation, see e.g. eq. (4.1) in {{ :teaching:ft1920:vl:cs:files:magma.pdf |Hovhannisyan et al, 2016}}. Summarize the convergence results of the more recent work {{ :teaching:ft1920:vl:cs:files:ergodicconvergencecp.pdf |Chambolle-Pock, 2016}}. | To apply FISTA you need to consider the Lagrangian formulation, see e.g. eq. (4.1) in {{ :teaching:ft1920:vl:cs:files:magma.pdf |Hovhannisyan et al, 2016}}. Summarize the convergence results of the more recent work {{ :teaching:ft1920:vl:cs:files:ergodicconvergencecp.pdf |Chambolle-Pock, 2016}}. |