Slashdot Mirror
Photoshop CS5's Showpiece — Content-Aware Fill
Barence writes "If you're looking for reasons to upgrade to Photoshop CS5 when it arrives, a new demo video might just persuade you. Narrated by Bryan O'Neil-Hughes, a product manager on the Photoshop team, the video shows the new content-aware fill tool, which has the potential to revolutionise the way you clean up photos. If you're not happy with an item in your picture, select it, delete it, and Photoshop will analyse the surrounding area and plug the gap as if it never existed."
Stalin would have just loved that content-aware fill tool.....
Cool! Maybe I can use it to create a picture of me with a girlfriend so my mom will get off my back!
I wouldn't call this an ad. This is legitimately really fucking cool.
"16MB (fuck off, MiB fascists)" - The Mighty Buzzard
Fox News and the Texas board of Education.
When I took a basic digital image processing class in college, we learned some of the basics of image processing (really more segementation than anything). To think about the complexity and sheer power needed to do this...just blows my mind. Truly impressive.
Photoshop currently sells at a "lightweight" $700. How many photos would I have to edit to make that cost effective? It entered the land of exclusive pro tool years ago.
Content-aware fill is different from content-aware resizing. Content-aware fill is basically a clone brush that automatically decides how to make the cloned area match into its surroundings instead of just copying and pasting content.
Umm, if GIMP was ever able to interpret fill EDGES! Did you even watch the video? It is holy impressive how it filled in the edge gaps of this panorama image at the end of the video. Just, WOW! I WANT TO SEE THE CODE THAT DOES THIS SO BAD!
I call Shenanigans. Someone has released this video a few days early.
If anything, go to around 4 min, and watch till the end (only a minute left). This is groundbreaking. Insane. I hate it, gonna put me out of my job.
Photoshop has had that capability natively (ie, not requiring a plugin) since CS4, this is the ability to select an object in an image - litter on a lawn was the example given in the article - and replace its former location in the image with content derived from the surrounding areas. Basically it's like an intelligent, automated version of the Clone Brush tool on steroids.
UNIX? They're not even circumcised! Savages!
Sounds like something Google Street View could use to remove people from their views and make them more acceptable.
Ex wife, there you go
Great for folks who go through relationships like Octomom goes through diapers.
I edit photos for my sisters business (she's terrible at taking photos of her products, always cutting bits off I have to fill out and work around) and I can safely say this feature would have saved around 60% of my editing time. That's a conservative estimate- probable around 75%.
Can't speak for you guys, but I'm sold.
Fox News and the Texas board of Education.
You know what's really funny? When those two collide.
My work here is dung.
I wish there was a paper on the core algorithms behind it (cmon Adobe, SIGGRAPH), but I could see why Adobe may be sitting on every aspect of this tool; because it sure seems to bring some real photoshop wizardry to the common user. It was really an example of "delete this thing" and it just works. Takes a common complex task and massively simplifies it. One of the most impressive marketing demos I've seen in a while.
Sure, there are some cases in which I doubt it works, but from what I could see, it seemed to have some vision and perceptual rules built in to guide how to fill in the deleted area. And frankly, it's a feature that for professionals, makes the price tag for the upgrade worth it. For some tasks, it'd pay for itself in labor alone. What would take a expert hours to do, this could do in minutes. If I was Adobe, I'd seriously consider taking this and make a Photoshop Elements Extended Edition (or whatever) and add about 79-99 bucks to this price for this feature alone. Arguably, it'd be worth it for many.
Next up is the "CSI Enhance" tool. Take a photo of 10x10 pixels, and make it a perfect 2MP image.
No-one else dubious as to the actual quality of these touchups? I'm definitely not convinced when all that's available is a SD quality video that's been through YouTube's encoding, since it barely shows anything more than the basic colours and shapes are there.
I wonder what sort of a fig leaf it will use to plug the gap in the goatse photo...
I want to play Free Market with a drowning Libertarian.
Guillermo Sapiro developed the first remarkable inpainting techniques. I'm sure Adobe uses similar techniques since I remember some joint research being done with the U while I was there.
http://mountains.ece.umn.edu/~guille/inpainting.htm
Go Gophers!
I like to pirate each Adobe CS just for the sake of having such expensive software. XD
Wasn't this core technology discussed on Slashdot a number of years back? If you google "Seam Carving" you'll find some nice wikipedia articles that discuss content-aware image resizing. This may be a variant on the same technology, and i actually doubt that this is an early release of an April Fool's Day joke (no matter how Star Trek this technology seems).
Liquid rescale is an implementation of the Seam Carving technology which was incorporated into Photoshop CS4 as a feature titled Content Aware Scale.
This new feature comes from an algorithm titled PatchMatch which was presented at SIGGRAPH 2009:
http://www.cs.princeton.edu/gfx/pubs/Barnes_2009_PAR/index.php
I remember reading about folks at Carnegie Mellon building this tool a few years ago.. Sure enough, there's a slashdot article.
I suppose the photoshop tool only pulls from material in the photo you're editing, and not from outside sources.. so it's a bit different. But the goals/results are pretty similar, it seems.
Mod parent down for not RTFA; this has nothing to do with liquid rescale.
http://cimg.sourceforge.net/greycstoration/demonstration.shtml
(Image Inpainting section)
Why is this idiot modded informative? He should be modded off-topic and clearly didn't read the article... I guess neither did the mods.
isnt the way with the software companies these days ? fill a software with innumerable features even professionals will rarely use, and ask $60-100 for 20 to 30 functions/features that people will use, because you also put there 180 or so ones that noone will use.
that's why software is being pirated. noone wants to shell out $60 for 200 functions 20 of which they will use from time to time.
Read radical news here
I for one welcome our automatic fake nude image generating overlord.
Boobies!
Welcome my "getting modded Insightful for not RTFA and spreading bullshit" overlords at Slashdot.
Content-aware scaling has been included in Photoshop since CS4.
But this is no scaling, it's filling. The first 3 minutes of the video are not so interesting (nothing one could not do with a clone stamp in 2 seconds), but the last 2 minutes are breath-taking.
GREYCstoration or liquidrescale don't even come close.
Now watch for FOSS fanboys claiming that this feature already exists via some obscure plugin to Gimp or whatnot. Of course, it won't even be close to the real thing. I didn't check, but someone has probably already made this post.
Yet another example of innovation in a commercial product, now watch how open source clones struggle over the next couple of years to make a 1:1 copy and eventually either failing due to the technical difficulty or making it horribly user-unfriendly in the process. If the former, they will claim "no one needs this feature" and if the latter, "shut up and read the README in the root of the tarball".
Expecting flamebait so posting Anon.
It'll be a great "mother-in-law" tool for family photos.
"This post is an artistic work of fiction and falsehood. Only a fool would take anything posted here as fact."
...on the iPad? ;)
I wouldn't call this an ad. This is legitimately really fucking cool.
So is the Mr. Clean Magic Eraser...
Did you watch it? This is really cool technology.
Frankly if I'm not coming to Slashdot to see the latest and greatest toys, technology, and, yes, products then what is the point...
So easy, a CAVE man could do it. Oh, wait, what?
Using Photoshop to remove lens flares? Oh! Brave new world!
-Peter
Similar algorithms have been around for a LONG time.
http://gmic.sourceforge.net/gimp.shtml
http://cimg.sourceforge.net/greycstoration/demonstration.shtml
http://tech.slashdot.org/article.pl?sid=07/08/09/131207
I see a 12 megapixel image in hand of a before and after and not a tiny less than 400 pixel overcompressed youtube video.
I have seen this automatic stuff before and when you look carefully at it it's not very clean unless you re-sample down to 1/4 the resolution or go small for web use.. it's never clean enough to print out at 11X17 or larger.
Do not look at laser with remaining good eye.
I once did a writing contract at Adobe. You know how when you pose for an ID photo, they put you in front of a curtain or something to hide the background? When I got my Adobe badge the security guy just posed me against a regular wall, then Photoshopped the wall out of the picture!
I'd like to see this perform on a patterned background. It's one thing to fill in a gradient, quite another to reproduce a pattern and repeat it properly.
It's not the same. Liquid Rescale moves the pieces by rescaling around them. This actually replaces just the exact area. I think it's quite a bit more useful, but in different ways.
"If you make people think they're thinking, they'll love you; But if you really make them think, they'll hate you." - DM
I can't wait until a crime show gets ahold of this.
"Delete that wall and see what is behind it. Enhance. Enhance. Enhance."
or else!
Yeah, and cars have already had the native ability to drive, turn and stop for a century. The DARPA Grand Challenge isn't really adding anything new.
Those robotic cars are basically just intelligent automated versions of cars, on steroids.
Just because it happens in software does not make it trivial.
Porquoi?
Includes more detail about the algo
- Developed with researches at Princeton
- Demo'd at SIGGRAPH in Aug. 2009
- Old spot-healing tool tried to find one match for the hole; new tool copies multiple patches from the surrounding BG to fit into the hole, as well as finding & copying surrounding patterns
http://www.youtube.com/watch?v=S9vbHRcrbdQ&feature=related
The English word fart is one of the oldest words in the English vocabulary.
Imagine using this for cutting stuff like your penis and then using the content fill tool(tm). People could compete on who would get the largest estimate ! Or nerds can finally remove their pimples from their faces easily, hurray adobe!! I currently suffer sleep deprivation.
So, what about Resynthesizer?
Ezekiel 23:20
Get the resynth plugin:
http://www.logarithmic.net/pfh/resynthesizer
These are graph-cut or similar algorithms. There are several free alternatives which have been out there for years. Two spring straight to mind - the resynthesizer plugin for the GIMP and GREYCStoration image inpainting.
CS5 seems to have made this easier to use but the functionality has existed for ages.
Cheers,
Toby Haynes
Anything I post is strictly my own thoughts and doesn't necessarily have anything to do with the opinions of IBM.
That capability has been available for a while in the Gimp as part of the Resynth plugin:
http://www.logarithmic.net/pfh/resynthesizer
It lets you resynthesize a texture, fill in a selection with surrounding content, and synthesize images "in the theme" of another image.
Basically it's like an intelligent, automated version of the Clone Brush tool on steroids.
There, I've highlighted for you why it's a big thing. I can't quite tell if you were unable to determine this fact because you're a moron or if you have some kind of language processing brain defect that makes you unable to consciously pick out salient points of information.
It fills in detail from the surroundings. I want one that fills in detail from the underlying layer. For example when you delete clothing.
Nullius in verba
a product for 'home user market' is sold from $100. you are gonna buy one, and 3-4 times a year you are going to rotate, crop and apply a few effects to your kids' photos.
no wonder it is being pirated.
Read radical news here
I found a higher-resolution (cropped) copy of the panorama used in the video:
http://www.scottkelby.com/blog/wp-content/uploads/2009/02/7_deathvalley.jpg
(According to the description here, photomerge, dodge, burn, and sponge were used, but basically only the colours have been changed.)
Composite I made of image from YouTube video + image from the blog: here
Alexander Peter Kristopeit bought his basement from his mommy for one dollar.
Uh, no. That is the patch tool. If you'd read the article and watched the video you'd know that this is far more complex than that.
The content aware healing brush and fill are taking hundreds of small sections from neighboring areas of the image and piecing them in.
Very kewl technology for its intended purpose. But also a great too for removing watermarks from images now.
Why is Adobe supporting the pirating of images... [sarcasim]
I'm usually among the first in line to call Slashdot on its thinly-disguised slashvertisements, but this goes beyond product upgrade and into the realm of William Gibson novel.
I'm kind of staggered just trying to wrap my head around the uses and implications...
The key thing is getting it not to chatter or flicker, which it probably will- as I doubt it will generate the exact same results frame to frame. Nevertheless- expect it to make matte painting, wire removal, etc a lot easier. I expect they'll use it to generate a quick starting point for clean plates, which will then be given further refinements and then composited in normally.
After watching the video and seeing obvious problems even at 360p, it seems unlikely it'd hold up at 2k without some love at least.
www.GrenadeHop.com
a few more years and they'll make up the whole picture.
I take it you didn't watch the video then?
I did watch the video. Content-aware fill does roughly much what Resynth does. All the examples in that video work in Resynth.
Looks like it has been photoshopped. Not real.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
Sounds like something Google Street View could use to remove people from their views and make them more acceptable.
Why would people become more acceptable if they are removed from Google Street View?
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
How well does it handle boobs?
I want this tool in Final Cut. It would take insanely long to process, but man oh man would it be useful.
Facebook is the new AOL
Very true. It just means that they'll be able to fund even more interesting research into it. :D
Eventually, I bet you'd be able to 'shop out the annoying drunk guy on the fringe of your family beach videos, but I expect that seeing it on the big screen (on in commercials) is only a few years away. Frighteningly, I predict that it could also be used for news manipulation -- hide those protesters, hide this pile of bodies, etc. Still, technology is still just a tool, and it's neat to see this sort of research.
Just "wow". Everyone who has spent tedious hours "fixing" some piece of "almost" perfect photography just fell off of their chairs.
I haven't bothered upgrading anything but InDesign in recent years - the old Photoshop (or even GIMP) was good enough. This is a reason to upgrade!
Enjoy life! This is not a dress rehearsal.
http://www.youtube.com/watch?v=dCr96ldClz4&feature=related
Same stuff except with real time video. Amazing.
Stop with the ads! That's like the 4th one today.
News Flash: Most of the 'news for nerds' we care about involves products we can go buy. You're going to have to try another meme to earn a mod point.
"I like to lick butts!" by MobileTatsu-NJG (#32700246) (Score:5, Informative)
I'm not into graphics much, but this seems similar to what Liquid Rescale can do.
http://liquidrescale.wikidot.com/en:examples
At the 3:45 mark in the first video they show removing elements.
Also ... it's only remotely on topic, but your FX reel (on the site you have in your sig) is pretty cool.
The skin effects in the Iron Man footage are neat, but the manufactured buildings and landscape composites are fantastic examples of the relative realism of manipulated videos. I really liked the live comparisons of "original" and "retouched" bits, because it was often fairly hard to know what I was looking for (esp on the digital makeup effects in the Iron Man clip). I'm sure it would have been even more interesting had I watched the higher resolution videos. Awesome!
And is now lightly sprinkled with a tasty cocktail of viruses and malware. It will serve you right when your thieving ways are rewarded with an empty bank account. If you want free, do it the responsible way and go open source!
As a photographer and retoucher, this looks to be something worth while. As others have mentioned earlier, there are probably some things that will need to be taken care of around edges and stuff, but still, this could save some serious time and headaches. I think the next step for Adobe should be a "beautify" filter, and then I can stop thinking to myself "Oh god, what have I gotten myself into," when I meet some of my clients for the first time.
But that's for cleaning things. You're not supposed to sniff it.
This looks like something from CSI. Couple that with infinite zoom and 4 huge monitors in a dark room and you could reconstruct court admissible evidence with ease!
i can't wait till someone is convicted of murder because content aware fill draws them in a picture taken next to an alley where the crime took place!
It seems to have intelligent pattern awareness. I'd love to plug something like an image of the rosetta stone into it and have it fill in the words in the correct languages of the broken off parts. Or maybe use it to translate documents from one language to another based on it's pattern recognition (scanned pages, after all, are just complex patterns of words - nouns, verbs and adjectives)
moox. for a new generation.
I quickly grabbed some screen shots of the images used in the demo and ran them through the GIMP "Smart remove selection" resynthesizer plugin (http://logarithmic.net/pfh/resynthesizer). The results were pretty good.
I uploaded the screen capture to http://tinypic.com/player.php?v=1sgw88&s=5
It's a very cool GIMP plugin for some things, but...
This is my source image:
http://s3.images.com/huge.28.142421.JPG
I want to remove the lady on the right, so I select her:
http://img714.imageshack.us/img714/1346/resynthesizerselection.jpg
And then, per the Resynthesizer page's recommendations, I use "Script-Fu/Enhance/Smart remove selection..."
http://img121.imageshack.us/img121/228/resynthesizerresultradi.jpg
Oh dear.
Anybody with access to the Photoshop beta feature want to give that image a stab? For all I know it fails just as spectacularly - but from the research it's based on, I highly suspect it'll fare better.
He's not saying the clonestamp has been in there since CS4 and this just automated it. The clonestamp has been in Photoshop since the dawn of time. He's saying this exact tool has been in there since the last version, so it's not technically "new". And robotics cars ARE basically intelligent automated cars, that's the whole damn point. Just because you somehow think writing "intelligent, automated" software is trivial, doesn't mean that any time somebody uses those words, they must be dismissing it as well...
ASCII stupid question, get a stupid ANSI
Now your ex-girlfriend won't even need to tear holes off her group photos.
At least I can finally fix those scary faces of lurkers in pics for "when you see it, you'll shit bricks" pics.
Freedom of "speech" lawsuits coming in 5...4...3...2...
Fill in the Blanks: Using Math to Turn Lo-Res Datasets into Hi-Res Samples
Considering that the regular image clone tool did that in one click (well, not counting tracing a selection and setting the clone source), I don’t think it’s a very good test image.
(I used GIMP, not PS, if it makes a difference.)
Alexander Peter Kristopeit bought his basement from his mommy for one dollar.
I wouldn't call this an ad. This is legitimately really fucking cool.
So is the Mr. Clean Magic Eraser...
Come on... that's just cheap melamine foam with some extraordinary marketing and retail markup.
Want to improve your Karma? Instead of "Post Anonymously", try the "Post Humously" option.
After Effects CS5:
http://www.youtube.com/watch?v=dCr96ldClz4&feature=related
You picked it out. :p
Alexander Peter Kristopeit bought his basement from his mommy for one dollar.
That's pretty fucking cool.
/* No Comment */
darn blockquote fail :)
I know I picked it out - what about it?
I didn't pick it out to specifically make Resynthesizer fail - it's image #2 on images.google.com for 'person in field' (sans quotes).
For an example that does work with Resynthesizer, try:
http://media-cdn.tripadvisor.com/media/photo-s/01/0f/33/e2/so-cool.jpg
Select the top-left dark thing, run the Resynthesizer script-fu - voila... dark thing removed, and sky filled in pretty well.
The problem is that this is entirely hit-or-miss.. and it's far more often miss than hit.. and then -when- it is a miss, it's a spectacular miss (as in that person-in-field image).
I'd just fill the selection with stars, hydrogen dust, and asteroids. That is, after all, contextual to everything and is never completely wrong.
Well, there's really nothing stopping people who want to manipulate the news to do so now- remember that photo Iran used to show it launched a bunch of missiles- but it turns out 2 of them were photoshopped in?
We commonly now do things like crowd removal, crowd duplication, removing garbage (AKA bodies?) from shots- and the software is all desktop based that will run okay on cheap PCs. The difference is the skill that accomplishes those things is still based in people- people who for the most part are based in Australia, US, UK, Europe.
The up and coming VFX regions are India, China and the Indochina area. Right now they're used for cheap basic film tasks; but in 10 years they'll have the technique and expertise down and a few shops there could easily match what the average shop here puts out.
With stuff like this in video, it's probably a little cause for concern but even if the technology is 5 years away, it'll still be detectable in 5 years if they don't have the talent to clean it up normally.
Now, 20 years from now.. Oh boy.
www.GrenadeHop.com
I think your parent was suggesting that the smoothing algorithm could take into account more than one frame at a time (hence no flickering or chattering).
Thanks man! Truth be told, I really need to remove that Love Guru building shot from the reel- it's only there to demonstrate that I know(knew?) Massive, a crowd generating software. It's not a very good example of compositing :(
Before and Afters are really critical to see what was done in a shot and how well it was done- but most VFX shops/Movie studios refuse to give befores to artists.. and then those same shops/studios when hiring ask for reels with before/afters- how the heck is that supposed to work?
Since that reel (I need to update it.. that's like mid 08), I've done work on Benjamin Button, Star Trek, Night at the Museum 2 and a little movie called Cabin in the Woods; but I won't have befores for any of those :(
www.GrenadeHop.com
That's not AE CS5, that's a student project demonstrated at Siggraph a year(2?) ago. I'm sure Adobe would love the technology, and might have made moves to buy it, but that video is straight from the student demo.
www.GrenadeHop.com
Watch the video linked to by the parent poster. It's arguably even more impressive than Adobe's demo...
-- If you try to fail and succeed, which have you done? - Uli's moose
just curious: do you read all of the comments you respond to, or just the first few words; and then log in with your "moderator" account to mod yourself up a bit?
sensitive file systems yet?
How many releases do you need to fix whatever retarded things you're doing to filenames to make case sensitivity a freaking issue?
Persistent Volume manager for Kubernetes - https://github.com/dwimsey/openshift-pvmanager
This is what I got very quickly with Alien Skin's Image Doctor: http://img530.imageshack.us/img530/1785/edited28142421.jpg
I thought about purchasing Photoshop for this feature, but it turns out our productive geeks out there have already implemented a plugin for the GIMP.
http://liquidrescale.wikidot.com/en:examples
I've used it. It works!
I suspect "Content Aware" means it *knows* it's fixing a lawn, or sky. The fix on the panaramic sky inserted a new cloud on the left! Cool, but will it know to insert a new zoo animal when you remove the concession stand in the zoo? The comment about seeing a detailed result is right-on, too; these things *often* look much better from a distance.
I've been doing professional level photo retouching for almost 11 years. Although this looks like just a better version of the healing tool/Patch Tool with a new flashy name, you can still even in the small crappy video see blurred circles left behind on the edges of the grass and shadows. Tree removal? Looks ok on a 480p, 3 inch wide video - print that image 20x30 and let me see how seamless and unnoticeable that removal is.
I like the direction it's going - although I highly doubt it's there yet...and it really wouldn't have taken that long to remove that tree manually using perfected techniques or the poles/shadows for that matter. How does it look removing a gouge/scratch from someones plaid shirt in a photo restoration? My guess would be - not so good.
Another useful "advanced" plugin is LiquidRescale:
http://liquidrescale.wikidot.com/
It implements seam carving. It's not yet available as an Ubuntu package, but hopefully someone will package it up.
True, it does make it easier to do now. However, it was always possible, just more work.
As usual, you take the good of progress with the bad. I've never considered watermarks to be a good way of protecting images. I watermark mine, but I usually want my name to be very visible, so when possible I try to put the watermark in the most solid colored corner. I know it's going to make it easier to clone out, but it's also easier to read. I take a lesson from the many complains of DRM. You can't stop the bad people, so don't focus on them....focus on your actual (or potential) customers. If someone is bent on stealing my image, I really can't stop them, so I try not to worry about it so much (and deal with it on a one-off basis after the fact if I find out). Instead, I'll concern myself more with making my watermark easier to read so people can more easily see who took the photo (provided someone doesn't remove the watermark, but again, nothing much I can do about it, so I don't worry).
Aside from that, your best mechanism of protection of your work (at least from my point of view as a part-time professional photographer) is to downsample what you post online. At least it gives me some sort of protection that dishonest people can't deal with. At least until this sort of smart fill is adapted in the next version or two of photoshop to become a smart resample (though that's got it's limits, because important details and textures may be lost in the downsample and not recoverable).
Now people can get rid of all those annoying watermarks on images that people don't want you to use!
Resynthesizer worked pretty good for me. see http://www.mesamike.org/misc/resynthesize-result.jpg.
Microsoft Digital Imaging 9.0 and 10.0 has this ability built in. The tool is called "smart erase", and fills in a selected area with patterns from the surrounding photo. I've had it match the pinstripes on my shirt in photos. This is a great product that has unfortunately been discontinued by Microsoft much like Flight Simulator. Even when the software was first released, it did not retail for over $50.00, much more bang for your buck than photoshop.
I'm very impressed but in the last picture (the landscape one) i think that is really impossible. The process even put a flying bird in the upper left corner. It doesn't feel like that is even possible.
This is what I get using the plugin on its own: http://shishnet.org/ufufuf/resynth2.jpg
Do note that the script-fu wrapper works better for larger images, which this isn't
Also, the example from the video, done with gimp instead, the results are pretty similar (IMO, better, but I'm pretty sure that the "improvements" are just luck): http://shishnet.org/ufufuf/panorama-synth.png
Having been using the resynthesizer for years, I've developed a knack for which source images will work well and which won't, and the thing that struck me about that video was that the source images are pretty much ideal conditions -- I'll be impressed when they can get good results on the images that aren't so clean :-)
I mod down anyone who says "I will be modded down for this", regardless of the rest of their comment
sample and fix, total fix time, about 3 seconds, two operations. Which is not to say that the Photoshop tool isn't cool; it is. But these aren't difficult tasks unless you simply have little skill at image editing. The video talks about those tiny, super-easy edits of the scene with the tree and bench as taking "all day"... that's a spit coffee, LOL moment, right there. Total fixup time for that image, same issues addressed... maybe five minutes or so. The best part of the demo was the outward fix of the missing sky on the panorama; that would take at least five minutes by itself with the tools I have (to do it as nicely.)
I'm sure the new CS will be great. All of Adobe's releases are great. The only thing better than CS is the marketing behind it. :)
I've fallen off your lawn, and I can't get up.
Only if you don't really care what the surface looks like afterwards. You almost always end up with a change in the reflectivity of the surface afterwards. I put a couple of shiny spots on my flat interior walls before realizing what was happening. We decided to go back to using 409, paper towels, and elbow grease.
This is my sig. There are many like it but this one is mine.
Just tried expanding a panorama as in the demo too, and resynthesizer does that pretty well too (please ignore the fact that the source material is REALLY shittily exposed, it was shot on a mobile phone with no manual exposure mode :( ):
original
expanded
I mod down anyone who says "I will be modded down for this", regardless of the rest of their comment
Now i can remove every baby jesus from every manger.
Filled using the "Content-Aware Fill" feature:
http://img245.imageshack.us/img245/6178/huge28142421out.png
Cool - that does look a lot better; still not as good as the fellow-commenter's quick clone brush (rubber stamp yadda) effort which I think would be a more natural path for a texture synthesis plugin to take, but a lot better than what I kept getting.
So - obvious question: workflow / parameters? Obviously I missed something (I did play with the parameters, the sky was pretty much stuck in there). Just wanting to see if I missed something (non-)obvious that might make Resynthesizer work on more images I've had trouble with in the past.
I'd still imagine Adobe's new toy would work better (it has its own flaws - which the presenter is quick to scroll out of view), but I'll gladly stand corrected on Resynthesizer's typical results.
obvious question: workflow / parameters?
Generally I use the foreground select tool* to select the smallest area to cover the object, then grow selection by a few pixels so that none of the object's edges are poking out and confusing it, then filters - map - resynthesize (ie, use the plugin rather than the script), and have the "tilable" options disabled since they tend to grab samples from the opposite edge of the image (if I want a tilable image, I'll use the tiling filter separately...)
Probably the biggest factor for simple success is to have the object you want to remove be on its own (surrounded on all sides by similar textures) -- if it isn't, then you need to do things the long way -- eg, if you want to remove the leftmost wheelchair from this image, and you want it to be replaced by grass when three of its borders are touching non-grass, then you'll find that it ends up somewhat messy since it attempts to merge four different edge textures. In this case you'll need to copy a section of your desired fill texture (ie, a rectangle of pure grass) into a separate image (specifically, a single layer image with no transparency); then on the original image select the object to remove, open resynthesizer, and select the "fill texture" image as the texture source; this way the generated texture will both match the surroundings of the original as much as possible, while being filled with the "surroundings" that you've specifically chosen. Having taken a sample of "pure grass" and a sample of "pure stone", then removing the top and bottom halves of the wheelchair with each respectively, the results are nicer. (with the exception that the first two images were produced with a mouse and twenty minutes of careful selecting, and the final one was 5 minutes work with a laptop nipple, so there are still some bits of wheelchair poking out of the sides...)
Incidentally, does photoshop have SIOX yet? Having the features "vaguely scribble in the general area of an object to have the object selected precisely" and "automatically and realistically remove a selection" could potentially combine to form "one-click realistic object removal" \o/
I mod down anyone who says "I will be modded down for this", regardless of the rest of their comment
As much as I like open source, I just downloaded Resynthesizer and it was very much hit and miss. I did as some people suggested and took screen shots of the YouTube video and tried to reproduce things in the GIMP. I was able to very successfully remove the tree from the sky of the first shot, but when I tried removing the trash and the poles, I got weird sky colored patches in the grass. The desert road got replaced with sky. And on the last photo, Resynthesizer did a very decent job on the clouds, but filled in part below the mountains with sky. I tried this with several of my own photos and consistently Resynthesizer seemed to show a bias towards filling everything with sky.
So, I am lead to conclude that while in principle Resynthesizer might be useful, in its current implementation it is very buggy and gives very unpredictable results. If Context-Aware Fill makes it into a release of Photoshop, I'm sure Adobe will have spent a lot of time and effort going the last mile and eliminating all the wonky corner cases.
It's *possible* that the video showcased an one in a million image for which this technology works particularly well, but I don't think Adobe would ever release a product which makes obvious mistakes most of the times. If they are seriously planning to release this, it's likely they've done considerable tuning to the algorithm and have gotten it to behave in most cases.
Bit early for April Fool's Day, but the last image proves it. There is no way in hell that any program would take the upward curve of the hill in the lower-right background and turn it into a DOWNWARD curve. That's something that you could only possibly know to do if you had an original image where the hill curved downward. Not to mention all the cloud detail that simply does not exist anywhere in the panorama.
Clearly all they've done is *added* a bunch of things to images (or in the case of the 'panorama', just deleted part of it to fake a panorama), then "used the tool" to "remove" them.
I am scientifically inaccurate.
Next time, Hurricane survivors won't be accused of looting, they'll simply be edited out of the shot and the news will say that everyone escaped unharmed. :-o
"Smart Carver - Seamlessly remove objects from an image with professional quality. The Smart Carver also lets you expand or contract objects without distortion."
This tool may not be quite as sophisticated but it's cheaper than CS5 Photoshop.
-Eric
Another option is G'MIC and it's inpaint feature (unfortunately the algorithm is not able to reconstruct textures).
It's also not that new. My MATLAB lecturer demonstrated the same thing with static pictures last week and explained how it works. It doesn't too difficult to apply that to a video, if you just add another dimension and compare frames to identify scene changes.
Most human behaviour can be explained in terms of identity.
and because, wisdom and rationale and logic are only conferred to people who go to a scholastic, rigid education institution right ? this is the reason why we had innumerable numbers of self-schooled, or even unschooled geniuses back in 19th century ....
sarcasm off.
turning in a nice profit does not mean you cant turn in even more.
you have to fit your business model with the behavior of the people and times. not only your price.
Read radical news here
if you just add another dimension and compare frames to identify scene changes.
Which, incidentally, is exactly what video encoding with compression does. Insert a single key frame with edits, and you may have changed several seconds worth of video with no need to actually edit them.
Alexander Peter Kristopeit bought his basement from his mommy for one dollar.
VERY interesting, thank you. My wife loves this but it's ridiculously $$$ for what it is (A small piece of BASF Basotect V 3012 foam). I just tried looking for suppliers in Ontario and haven't found anything yet. I also just phoned BASF Canada and they don't supply or sell finished products.
No sign of melamine foam or Basotect on fleabay either.
I wonder when I can get it.
Tom...
Yup, agreed. That's some seriously nice bit of tool there. I've skipped CS3 and CS4, looks like CS5 is the one to get ;-)
Dave
Our lives begin to end the day we become silent about things that matter. --Martin Luther King Jr.
Try here: http://www.alibaba.com/trade/search?SearchText=melamine+foam
You may need to order it by the container load, but that's where I'd start.
Want to improve your Karma? Instead of "Post Anonymously", try the "Post Humously" option.
I would think video, being multiple still images, would actually be easier to do. As an object moves from frame to frame, it would reveal more background that could be used in other frames.
Umm, of course if the background is moving or changing a lot, that could make it harder!
Uncrop those old photos! I always wondered what my great-great-grandmothers only half visible-in-an-old-photograph dog really looked like.
http://www.youtube.com/watch?v=KUFkb0d1kbU
Some even better examples of this awesome content-aware feature:
http://www.youtube.com/watch?v=3ScWu7pG7r0
So amazing!
You can do this now using Mokey from Imagineer Systems. It patches together an idea of what an occluded background plane looks like (similar to CS5's Content-Aware Fill), but then goes on to track that plane over the occuluding object, thereby removing it. Here's a quick demo: http://vimeo.com/7533880
Do you actually read the comments you snark at anonymously?
Porquoi?
I've taken a shot at it, in the beta version:
http://lh5.ggpht.com/_6FsRfRNHJ8Q/S7L4e-cb2MI/AAAAAAAAC4I/fl6C3dtl07g/huge.28.142421%5B1%5D.jpg
Seems pretty good to me for 30 sec work :)