Denoising, deconvolution and computed tomography using total variation penalty

[@levant_noir]

Good work Michael! Very few people can be happy and smiling while talking about matrices :)

[@raviprakash1278]

Thank you very much sir for uploading this. This helped me a lot.

[@charlesemekaonyi8158]

Thank You. This is quite an insight. I will watch it again.

[@MohamedMohamed-br8lh]

Respect from Egypt 🇪🇬

[@rulisastra]

Wow, even i dont understand all of them, i'll say "the explanations are great!"

[@kmmertes]

I'm confused. For 1D signals you define TV as the L1-norm of the derivative of the signal. But for 2D images, you define TV as L2-norm of the gradient. Shouldn't TV for images be the L1-norm for dx/du1 + L1-norm for dx/du2? That is, TV = |dx/du1| + |dx/du2|. Or am I missing something?

[@MichaelZibulevsky]

You can call it L1-L2 norm: sum over pixels of L2 norms of the gradients

[@ItsMeMabu]

Thank you very much! Keep up the good work. :)

[@FatehBenMerzoug]

Very nice explanation, thank you

[@skovnats]

Спасибо! Thank you! Very useful. Artiom.

[@arlettecastillo9445]

There are more examples??!! :'(

[@DrPatZMusic]

nice tutorial!

[@MichaelZibulevsky]

thanks!

[@saparchitekt]

Interesna! Just show abstraction of real world to formula's and do not bother with solving formula's. Secondly don't bother with two dimensional index for x, just number all pixels in a single index!!.

[@MichaelZibulevsky]

:))