conan says...

AAAHHAHAHAHHAAAAHAAA HAHAHAHAAAA HAAAAAAHAAAAHAHAHA
!!!!!!!!11111oneoneoneeleven

I haven't seen CSI up until now, but now i'm sure this was the last bit i ever watched from it. what nonsens.

Structure says...

Best part is at 0:53. A low-resolution side view of her head turns into a perfect front view of her eye. If they can change the camera angles of a still image why not make it point at the killer? The CSI motto: "Computers is magical."

rottenseed says...

Not to mention the fact that they rotated the viewpoint of the eye...pretty hard to do from a 2-dimensional picture.

With that being said, the original CSI (Vegas) is awesome and they'd never do that.

HollywoodBob says...

I don't know what is worse, that they put this garbage into the shows or that people who watch really believe that these things can be done?

ponceleon says...

God you guys are just so over-critical.

Didn't you know that every time something happens you don't understand a wizard did it?

Well, a wizard, or Jesus.

potchi79 says...

Whatever camera that was could be used in diagnostic medicine. Who needs blood tests? Just take a picture of the patients tongue and zoom & enhance.

Almanildo says...

This reminds me of the (otherwise great) computer game Blade Runner. You had these camera images where you could barely see two people through the doorway. Then you zoomed in, and the camera sort of went through the door and rotated so that you could see their faces. It was so obvious that I almost wondered if you actually controlled a camera that flied around in the shop.

pierrekrahn says...

Thankfully he just happened to have a spoon with him that day, otherwise this crime might have never been solved.

It would have been funnier though if he accidentally grabbed his Swiss army knife by mistake!

vairetube says...

and whats crazier is the fledgling technologies and seperate softwares in the works that all eventually will be able to reprocess, extrapolate, and stitch (however accurately) images ... its not going to be science fiction forever, despite our current incomprehension. ... we have a program to resize pictures that lets you expand or decrease and not lose features, we have a program to make 2d into 3d.. theres photosynth...there's the program that *GASP* can render unseen portions of a face from a single picture....and other things, all of which ive seen on the sift.

but yea ..not yet. you damn cynics.

dont become the people who laughed and said the earth cant revolve around the sun!1 don't fuck wit science.

Xax says...

And those damned goofy-ass bleeps and bloops every time they so much as move the mouse... bad 50s sci-fi.

MaxWilder says...

>> ^vairetube:
but yea ..not yet. you damn cynics.


There is not, nor will there ever be, a computer program that can extrapolate a guy in a t-shirt carrying a basketball from less than ten pixels of a reflection in a person's eye.

In Blade Runner, he did some crazy extrapolating from a photograph. Not only was it set in the future, there was no telling what kind of resolution or future technology was used to actually take the photo, so it's not quite as laughable.

In a TV show set in the present, it is most certainly a Wall Banger, and perhaps even a Dethroning Moment of Suck.

Yes, I've been reading too much TV Tropes.

Sketch says...

No way, Xax! I totally want my copy of Photoshop to make those beeps every time it's trying to process something! That's rad!

kronosposeidon says...

I like how the Sift's *Geek Squad shows up every time for these videos. You guys realize that most people actually do believe this shit, don't you? Just because you and your geek friends know this is bullshit doesn't mean that Joe the plumber knows it. Of course you and your geek friends probably know plumbing better than Joe does too.

vairetube says...

it could Max... and it would be wrong...you mean accurately extrapolate. i believe you would be correct.. you can't make something from nothing.

i am willing to bet something equally as mind boggling, however, will be performed by a computer before i see the end of my days. its a matter of the data available to the machine and the machines themselves.

2d into 3d
http://www.youtube.com/watch?v=VuoljANz4EA

http://make3d.stanford.edu/

content aware resizing:
http://www.youtube.com/watch?v=qadw0BRKeMk

3D morphable face models
http://www.youtube.com/watch?v=nice6NYb_WA

facial recognition:
http://www.youtube.com/watch?v=P6I0k7UofR4
and related replacement
http://www.youtube.com/watch?v=MZMrNYSa9pI

photosynth:
http://www.youtube.com/watch?v=p16frKJLVi0

and finally... a realistic use of photosynth on CSI....
http://www.youtube.com/watch?v=0suot89qXY4

so maybe you can rejoin the fan club.


really what we're talking about here is insanely increasing the resolution of the original photo to begin with, so the manipulation software isnt the issue... maybe you have video technology that takes multiple levels/layers/scenes and can on demand get that data... thats more of an issue with the recording technology then analyzing.

but, you're right... NEVER EVER EVER EVER EVER NO WAY if we're talking about software.... or.... maybe there is and you don't know it yet... as illogical as it may seem.

yer basically right, but don't say no... think of a way.

http://www.hasselbladusa.com/news/hasselblad-announces-new-50-megapixel-camera-and-upcoming-645-sensor.aspx

--------> http://www.gigapxl.org/

good day!

[defunct] Drax says...

Now all he has to do is point the spoon in the direction of the suspect's house, and use a camera to magnify and x-ray the reflection on the spoon till they find him.

Cheap security camera should do it (resolution might not be too good though).

rychan says...

vairetube: Yes, there's a lot of cool research in computer vision, computer graphics, and computational photography. But it's pretty simple to say, in an information theoretical sense, that this CSI stuff is impossible.

Yes you could hallucinate plausible image content (super-resolution), but clearly that's not appropriate for a forensic setting.

But in this case we have a quantized, noisy, low resolution signal and the reflection in the eye could have been generated by any number of incident signals and it wouldn't make a difference in the recorded video signal. The information just isn't there.

ForgedReality says...

"The resolution isn't very good. However, I just managed to add about 45 gigabytes of information to a VGA CCTV still image by sheer MAGIC and IGNORANCE."

TV is for fat stay-at-home moms who don't know anything about what a computerbox typewriter is.

doogle says...

No-no! It's fine.

They reversed the image. Would totally work. The field of "corneal imaging" is really taking off. I'm getting a bachelor's degree of it as we speak.

Shepppard says...

I second the fact that vegas isn't as bad as this.

From what I can tell:

CSI Vegas: Sticks to actual possibilities to solving crime, doesn't do anything retarded for storylines, and has actually invented a couple new techniques from the show that are now used by CSI. Mostly seems to be based around the science behind the CSI.
Also, William Peterson kicked ass.

I actually have a random clip of the show of an analyst using the machine to break down the materials of..
something. can't remember.

http://www.videosift.com/video/CSI-Vegas-Do-your-dance

CSI NY: Seems to try to be off the wall, trying to re-invent certain things that could be used by CSI (As we see in this video) but they never actually work out, and it leaves you feeling both like a genius, and a retard for watching it.

CSI Miami: The entire show is based around sunglasses of justice.

Sagemind says...

Well, for One, Applaud them!!!

If they sew enough of this stuff into existence, maybe the actual criminals will be so paranoid when they commit crimes that they will think everything is a clue. They will be so busy trying to fix the forensic stuff, they'll forget about the obvious stuff and leave leave behind the "Big Obvious Clues" - Like a "leather glove" of something...

nach0s says...

>> ^Almanildo:
This reminds me of the (otherwise great) computer game Blade Runner. You had these camera images where you could barely see two people through the doorway. Then you zoomed in, and the camera sort of went through the door and rotated so that you could see their faces. It was so obvious that I almost wondered if you actually controlled a camera that flied around in the shop.


yea based on a scene from the movie, which works because it's THE FUTAREZ

At the 4:59 point in the link below, you can see a version of the scene.

http://www.youtube.com/watch?v=2MiDsD6N65Y&feature=related

harry says...

It worked better in Blade Runner. Because that was obviously set in a fantasy future world.

I like how they zoom in on the baseball in the end, just to make it even more obviously ridiculous.

mauz15 says...

>> ^HollywoodBob:
I don't know what is worse, that they put this garbage into the shows or that people who watch really believe that these things can be done?


There is something even worse. The college students who decide they 'want' to be in criminal justice because CSI is so kewl omg I want to be exactly like Sara Sidle!!!111!.

Almanildo says...

Ok, I know that scene from the movie, but I didn't remember the camera angle shift there

It's still way more obvious in the game. And yes, it's still ridiculous. I can take extreme zooms, but I can't take camera angle shifts.

siftbot says...

This published video has been declared non-functional; embed code must be fixed within 2 days or it will be sent to the dead pool - declared dead by pipp3355.

vaire2ube says...

>> ^vairetube:
5/2009
really what we're talking about here is insanely increasing the resolution of the original photo to begin with, so the manipulation software isnt the issue... maybe you have video technology that takes multiple levels/layers/scenes and can on demand get that data... thats more of an issue with the recording technology then analyzing.


http://www.petapixel.com/2010/05/07/omni-focus-camera-boasts-infinite-depth-of-field/

5/2010
Researchers at the University of Toronto have come up with a new video camera that can achieve infinite depth of field even when objects are immediately in front of the camera. What they did was stuff an array of video cameras into a single camera, with each camera focused at a different distance.


2009:
http://cg.cs.tsinghua.edu.cn/montage/main.htm
System that composes a realistic picture from a simple freehand sketch annotated with text labels. The composed picture is generated by seamlessly stitching several photographs in agreement with the sketch and text labels

Present:
http://science.slashdot.org/story/11/09/23/1325247/brain-imaging-reveals-the-movies-in-our-mind

"Brain activity evoked by the second set of clips was used to test the movie reconstruction algorithm. This was done by feeding 18 million seconds of random YouTube videos into the computer program so that it could predict the brain activity that each film clip would most likely evoke in each subject. Finally, the 100 clips that the computer program decided were most similar to the clip that the subject had probably seen were merged to produce a blurry yet continuous reconstruction of the original movie"


----


So maybe you just need to find a witness then you can read their minds! Blows the incredulity over CSI out of water.

vaire2ube says...

1.8 gigapixel ARGUS-IS. World's highest resolution video surviellience platform by DARPA.
1 million terabytes a day saved forever.

The ARGUS array is made up of several cameras and other types of imaging systems. The output of the imaging system is used to create extremely large, 1.8GP high-resolution mosaic images and video.

The U.S. Army, along with
Boeing, has developed and is preparing to deploy a new unmanned aircraft
called the “Hummingbird.” It’s is a VTOL-UAS (vertical take-off and
landing unmanned aerial system). Three of them are being deployed to
Afghanistan for a full year to survey and spy on Afghanistan from an
altitude of 20,000 feet with the ability to scan 25 square miles of
ground surface.

http://www.liveleak.com/view?i=e95_1359267780

the equivalent of 100 predator drones looking at one place AT ONCE ... hahah they stole my idea

vaire2ube says...

PWND.. THX FOR DREAMING... INFO ISNT THERE??? FIND A WAY TO MAKE IT THERE DURING CAPTURE. DONE. YAY.

rychan said:

vairetube: Yes, there's a lot of cool research in computer vision, computer graphics, and computational photography. But it's pretty simple to say, in an information theoretical sense, that this CSI stuff is impossible.

Yes you could hallucinate plausible image content (super-resolution), but clearly that's not appropriate for a forensic setting.

But in this case we have a quantized, noisy, low resolution signal and the reflection in the eye could have been generated by any number of incident signals and it wouldn't make a difference in the recorded video signal. The information just isn't there.

vaire2ube says...

the Illum is a "light-field" camera that captures information about the angle from which light arrives, allowing the device to allow users to adjust the perspective and focus of a picture after it's been captured.

who bookmarked this post:

Send this Article to a Friend



Separate multiple emails with a comma (,); limit 5 recipients






Your email has been sent successfully!

Manage this Video in Your Playlists

Beggar's Canyon