So how do I try it? Why do you have so many broken links, but no links to the covered material? I tried to find this site in google, but also unsuccessfully.
Literally needed this yesterday, but somehow ended up on almost a 1:1 copy of the demo website which doesn't work. Glad I decided to stay through the video, thanks a lot :]
Impressive video but i don't know/understand how to use it. I don't find any download possibility of any software on your website and i don't find any online solution like uploading a picture and downloading the enhanced result.
Most of the current AI. developments are all around human faces and animals. And this is just a very good example of that. But i am pretty sure, that they could also do the same for other stuff in the pictures, like identifying objects such as cars, buildings ect. to a restore level like this AI is doing. It's such a pain that it only works inside it's detection rectangle, even when the hair of the person outside of it, is there. Which is something like fabric, a textured surface with a lot of detail.
If you have a very blurry image, then it is impossible to know how the image will look sharp. This means that the resulting sharp image is (partly) made up. This means that the result do not show the actual person. It shows an image that would look like the blurry image if we do blur it. This might be cool if we want to create nice looking images. But, the result is no longer a photographic portrait. And, as it is not a photographic portrait, what value do it have to use this on old photos of old relatives? It is not them.
По моему фото в примерах перепутаны местами. Из хороших делали плохие. :) В названии "бесплатный", зато когда заходишь по ссылкам, всё становится очень даже платным.
No. You can not recreate information which was lost. The AI predicts faces based on their prior knowledge. The restored face will not look like the original face.
Doing aggressive or evil looking characters, I using CFPGAN to preserve the edge and sharpy feels, if with CodeFormer, it soften the looks of character faces and wash away the evil looks,made it look too kind which I really don't like.
Nope, i compared with github TencentARC GFPGAN) and results are worse or could be an alternative option, but non of them allows me to say "yeah, its next version". Sad, because its very fast compare to GFPGAN
Generally, they are good tools to have but having used both tools, I can say for certain, TencentARC GFPGAN smoothens face and in most cases loses the facial features like glasses - funny enough, if you test GFPGAN test images against CodeFormer, Codeformer produces a better quality overall without the smudge of pixel polish GFPGAN uses. GFPGAN: replicate.com/tencentarc/gfpgan Codeformer: replicate.com/sczhou/codeformer
Still needs a lot of work, even in non blurred photos it drastically alters the facial features. Sadly Remini is no longer free but it is light years ahead in terms of results from this one.
could this be used to enhance game faces in real time like it can do with video, or it does not do the video in real time and it take long time to render?
You can do that with Photoshop and it's also an old program that does it and with that the A.I thing is crap everyone is getting on board to collaborate with the Lie because something is being prepared for the future and the excuse will be the A.I (I know that There are things possible with AI but they are introducing a lot of unverifiable things and one thing is the software or another how the AI works
For the life of me, I can't figure out how to access this software on the GitHub site. Can you please explain in simple terms for idiots like me? 🙃 I'd really like to use it. Thank you.
![ARGENTINA vs. CHILE [3-0] | RESUMEN | ELIMINATORIAS SUDAMERICANAS | FECHA 7](http://i.ytimg.com/vi/De3AWmXzhuA/mqdefault.jpg)
So how do I try it? Why do you have so many broken links, but no links to the covered material? I tried to find this site in google, but also unsuccessfully.
CSI zoom enhance is becoming less impossible because of AI tools like this. It's insane and I love it
for movies is cool, for real life the final result might very different from the real appearance of the object.
Movies are 50 years after real life.
@@2010Edgars 😲
What do you mean exactly?
I wonder what is for you (the real life).
Your declaration, as it stands, makes no sense.
@@clementrodrigueHe means, tech in movies is 50 years in the future with respect to actuality. And what's realistically possible as of today.
Literally needed this yesterday, but somehow ended up on almost a 1:1 copy of the demo website which doesn't work. Glad I decided to stay through the video, thanks a lot :]
Glad I could help!
how did you get it to work?
"And how that works is simple" continues to show a highly complex slide which I have no idea what Im looking at. :D
ikr, I hate to know what he thinks is complex!!
Love how it saves so much cleanup time for A.I generated images and the wavy blurred images restoration, this is gold!!!
Finally those C.S.I scenes will be possible in rl
holy snap, how do we get to here in just 2 years? This is freaking INSANE
Impressive video but i don't know/understand how to use it. I don't find any download possibility of any software on your website and i don't find any online solution like uploading a picture and downloading the enhanced result.
In the past they said ENHANCE was just a hollywood trick, now its for real! LOL! So awesome!
The no one needs a good camera anymore. 😊
Most of the current AI. developments are all around human faces and animals. And this is just a very good example of that. But i am pretty sure, that they could also do the same for other stuff in the pictures, like identifying objects such as cars, buildings ect. to a restore level like this AI is doing. It's such a pain that it only works inside it's detection rectangle, even when the hair of the person outside of it, is there. Which is something like fabric, a textured surface with a lot of detail.
Yes, I am hoping for this as well.
How do i drop in an image in 'Code-Former' to fix it (from blur to clear) with clear information details ?
If you have a very blurry image, then it is impossible to know how the image will look sharp. This means that the resulting sharp image is (partly) made up. This means that the result do not show the actual person. It shows an image that would look like the blurry image if we do blur it. This might be cool if we want to create nice looking images. But, the result is no longer a photographic portrait. And, as it is not a photographic portrait, what value do it have to use this on old photos of old relatives? It is not them.
Good point, but considering the gear some of these pictures were taken with, how real was the original anyway?
"And how it works is simple" quantum mechanics of unknown origin show up
AskNK your channel is brillant.
Really good and informative content.
4:17 also need to do some work on eye detection, it changes the direction of the eyes a lot.
*Now use this on UFO footage, since that always gets taken through the worst cameras with the lowest-quality footage.* 😜
see ufo saudia arabia video
По моему фото в примерах перепутаны местами. Из хороших делали плохие. :)
В названии "бесплатный", зато когда заходишь по ссылкам, всё становится очень даже платным.
really waiting for img to 3d a.i to end user
I've just tested this software and it's amazing, to say the least. Thanks for the information.
How did you run it?
how do get it
Very nice overview
When I tried the Code-former it just spun around and spun around and spun around and never showed what the enhanced version of my photo was.
AMAZING, Tried the online 'demo' very impressed So how do you actually run this and make it do it's thing on a PC ?
This is crazy!
MAJOR!!!!!!
Cool, what about restoration of videos, can it be applied to videos?? For free?? 😮
Yea -although am yet to explore that, but it should work.
So when the FBI/CBI (Cops in general) ask the CCTV footage to be zoomed in, the culprits face will finally be clearly visible.
No. You can not recreate information which was lost. The AI predicts faces based on their prior knowledge. The restored face will not look like the original face.
lol if its going to be a blurry mess or one huge pixel don't expect for it to accurately depict the person...
@@Irrelavant__ That would be a footage instead of a photo, so I guess there'd be a lot of angles for the AI to analyze?
@@YOUnoobGER Exactly. Generating an upscaled human face based on educated guess doesn't mean it is the real face...
You have given the creator of tha appsan idea to sell their apps at high price'
How cool is that
CodeFormer is so smooth.
its intersing, its possible to test it?
Is it a stand online software or Photoshop presets-plugin or an online software?
This needs a tutorial follow up
OMG FINALLY ENHANCE!
thanks
Can it colorize black and white photos?
Thank you
It’s good invention n useful, thank you Sir
How to use it Can someone help me I didn't understand how to use it. Please.
Awesome video.
I'm subscribing !
thanks bro ... 🥰🥰🥰
Love frome Bangla Desh
How do I use the Wand software? Through GitHub?
My mind is totally blown
when theey make it easy to incorporate this into gaming engines it is absolutely gonna transfer gaming
Amazing and beautiful🤩
Doing aggressive or evil looking characters, I using CFPGAN to preserve the edge and sharpy feels, if with CodeFormer, it soften the looks of character faces and wash away the evil looks,made it look too kind which I really don't like.
I'm going to try with Goku and see what comes out :D
amazing work
Muito bom e como faz para conhecer e utilizar na prática gostaria de conhecer melhor
Amazing tool
Really Good
It seems it doesn't detect closed eyes and it makes eyes over the closed eye lids.
unbelievable
Can i use this locally on PC?
is this a 3D neural network architecture?
good tool, thank you.
Wow this tremendous
woooooow this is feature!!!
AMAZING !!! 🤯
amazing
O_o How can i use it?
The thing is of course cool, but my eye color has changed from brown to gray, what should I do?
veri beautiful AI
I hope in some years we ccan use this on an actual human vision
very good video thanks
The tool link wont open :(
Someone should use this on that Libby German video to see whether or not the guy on the bridge is identifiable as Richard Allen.
Nope, i compared with github TencentARC GFPGAN) and results are worse or could be an alternative option, but non of them allows me to say "yeah, its next version". Sad, because its very fast compare to GFPGAN
Generally, they are good tools to have but having used both tools, I can say for certain, TencentARC GFPGAN smoothens face and in most cases loses the facial features like glasses - funny enough, if you test GFPGAN test images against CodeFormer, Codeformer produces a better quality overall without the smudge of pixel polish GFPGAN uses.
GFPGAN: replicate.com/tencentarc/gfpgan
Codeformer: replicate.com/sczhou/codeformer
amazing what the website u testing on do u have link
Nice
great
Good job is the best
Very cool.😊👍
Wow!
Nice video
Still needs a lot of work, even in non blurred photos it drastically alters the facial features. Sadly Remini is no longer free but it is light years ahead in terms of results from this one.
could this be used to enhance game faces in real time like it can do with video, or it does not do the video in real time and it take long time to render?
magiccc
Name site Please
Como faço para poder adquirir essa ferramenta de restauração?
Someone needs to use this tool to finally show us what Bigfoot looks like.
Nice) 10X
Just waiting for this tech to work for bodies...
for no reason or anything, just maybe people out there may want that...
You can do that with Photoshop and it's also an old program that does it and with that the A.I thing is crap everyone is getting on board to collaborate with the Lie because something is being prepared for the future and the excuse will be the A.I (I know that There are things possible with AI but they are introducing a lot of unverifiable things and one thing is the software or another how the AI works
For the life of me, I can't figure out how to access this software on the GitHub site. Can you please explain in simple terms for idiots like me? 🙃 I'd really like to use it. Thank you.
CIA in real life, took a few years but it's now here 😂
I think you mean CSI. CIA has Already been in our lives, for far too long.
is this better than Topaz Gigapixel
Is it possible to run CodeFormer directly on my pc with ubuntu? If so, how?
que clase de brujeria es esta
Imagine combining this with face recognition software.
Yes, innocent people going to jail very soon.
Topaz Labs offers this now, for a price.
hey how do i use this
Bisakah anda membantu kami cara mengaplikasinnya dengan code former yang sudah terdownload?
But it seems just for face, but other details like clothes, decorations, and even the background that make whole image still look weird
Can I get the name of the App (iPhone), SW (MacOS) or Link?
how to try it?
Can this ai imrove videos but not necessarily with people?
how do we try this?
but the question is can it recover pixelated japanese xxx
Impressive technology. Companies, we will have to introduce periodically AI tools to our actual pipeline
Could you send me the link to the website?
I have just found a direct competitor to Two Minute Papers