Computer program reveals invisible motion in videos - Amazing!!!!!

Hi Guys,

Danny here. Check out these amazing videos that show the video magnification program using the Eulerian algorithm. Remember this technology first came out in 2010 I believe and then was kinda forgotten. Hope this does not happen this time round cause it so much easier to use now.

Okay, check this YouTube video created in 2012. Shows video of face .. She the blood flow, pulse, motion magnified...

2012 video showing face...

Next video. 2013. Shows many examples of how this program can be applied. Remember UFO, ghosts, politicians etc at the back of your mind... It shows in a 2.3 minute video the video filter setting .. It's really good

filter settings shown on videos of many things. love the one of the water and the heat coming of the pan fry up..

And this video explaining the technology skip to 1.25 for video examples....

explaining the technology video..

Guinea Pig Tiffany is the first rodent in the world to under go Eulerian Video Magnification! Cute!!!

Here is the video of Tiffany...

Tiffany the Rodent

Enjoy. Just a bit worried this open source software will be forgotten on ATS after a while. Pity there is no section on ATS to keep this as a permanent Topic.

Danny.

reply to post by Zer0megAlpha


You need MATLAB Compiler Runtime installed. Just download the one for your OS, and install.

Linux: people.csail.mit.edu...
Windows: people.csail.mit.edu...
MacOS: people.csail.mit.edu...

This is confusingly exciting.

Ive read the thread and watched all videos, admittedly im not a tech wizard so i would like to tilt my head in appreciation to those who have taken the time to test it and share here with us. Without your contributions I and probably others would never of known about this process.

Sending positive vibes your way!



I myself have a video i would like testing if someone would take up the task???
Theres nothing over exciting in my vid, its just some lights i filmed that didnt show up much on video.
Im curious to if they would show up when put thru this program.
My feelings are its just to poora quality but i wont know without asking.

Basically i filmed lights bouncing and dancing thru cloud cover, the 2 bright lights are street lights, the ones to focus on are the top right quarter of the screen. Please i know they are hard to see but they are there and if this programme isnt any use for my video, can some suggest a way of discerning the lights more?

heres the video

reply to post by GezinhoKiko


Your video, as it is now, is not good for this process, as it's too shaky, but I suppose that a stabilized version would work.

reply to post by ArMaP


Thanks for replying ArMaP

Can you suggest the easiest software to stabilize the video?
Its been bugging me since i filmed it, no one can see the flashing.
If you move your viewing angle around the screen they are barely visible, little flashes.
But with the naked eye it was quite a show.
If i can make them more visible/clearer then i wont be advised to go to specsavers lol (other opticians are available)

reply to post by GezinhoKiko


I have used a VirtualDub plugin called "Deshaker", with good results, but, for some reason, VirtualDub doesn't like Windows 8.1 that much, so I haven't used it lately.

Both are free.

reply to post by ArMaP

Thanks again ArMaP,
After 2 hours of trying to get the video onto virtualdub i have given up
I just dont understand what i should be doing, needs to be a something or other file "code"

Its all jargon to me!
never mind.



TECH 1 - GEZ 0

reply to post by GezinhoKiko


Sorry, I forgot that VirtualDub is not user friendly.

Took a quick look at the thread since I last was on and didn't see this. If it is already mentioned, sorry...

It hadn't occurred to me that this software could be valuable for this.

Source

To help filter out the phonies and fakes all trying to make a quick buck on something he believes in, Knapp is asking those vested in Bigfoot research to rely on technology built out of MIT to prove that the truth is out there.

Knapp, who runs a blog called “Bigfoot Crossroads,” a personal site with updates about all things Sasquatch, recently stumbled upon an invention created by students and researchers in the MIT Computer Science and Artificial Intelligence Lab in Cambridge in 2012, called “Eulerian Video Magnification.”

That makes a lot of sense. A guy in a fake Bigfoot suit should be easy to spot with this method while a genuine, living being should show movement of muscle mass beneath the skin as it moves. Way too many fakers are spitting out video. This could possibly end the old "Catch-22" of a video being either "Too Clear" to be true or "Too Grainy" to be of any value.

CornShucker
That makes a lot of sense. A guy in a fake Bigfoot suit should be easy to spot with this method while a genuine, living being should show movement of muscle mass beneath the skin as it moves.

Now we only need for Bigfoot to pose for a high definition, high quality video for some seconds, without moving.

ArMaP
Now we only need for Bigfoot to pose for a high definition, high quality video for some seconds, without moving.

So it only works with an object that is basically stationary? Guess I misunderstood, not the first time...

reply to post by GezinhoKiko


What type of camera was used make/model?

CornShucker
So it only works with an object that is basically stationary? Guess I misunderstood, not the first time...

It works better with high resolution, high quality videos, without too much compression (compression changes the way the image looks) and with stationary (or almost stationary) targets.

reply to post by wmd_2008


Hi wmd_2008,
I used a LG KS-360 mobile phone.
Pretty useless so i dont think theres anything that can be done to see the lights better.

cosmikDebris, a member posting in this thread came up with what I thought would be a great idea.

He suggested that someone use this new tech to analyze this video...



Anyone interested???