11.9: Computer Vision: Blob Tracking with Persistence - Processing Tutorial

In this Computer Vision video, I develop a "blob persistence" algorithm. I assign IDs to each detected blob and track its position over time.

Links discussed in this video:

Help us caption & translate this video!

[paladin]

This was amazing, I loved that this was forty minute so I had more time to take in with you . I still don't know exactly how this really works but I am getting hang of it and I did step with you so I really could take it in and it helped. Amazing idea btw xD

[danthemancasey]

Ha ha wow I gotta say this one had alotta ups, and alotta downs, but we kinda got there in the end! TBH I spent more time watching this video at .75x then any other in my life, and still had trouble keeping up typing as you went through it! Thank you for another challenging and informative video!

[elishmuel]

14:45 take your time buddy, soaking it all up!! Great stuff

[geoffwagner]

may actually be the most magical thing done with red cups to date

[ignaciogattarelli]

man, you're awesome! i always watch your videos! i'm still waiting for some library to connect my camera dslr to processing :(

[bastuebducloy]

hi Daniel ! i have a really important question: can we do the same thing like trackColor like you in the video in p5.js ? i like you're video ! continue like this!

[carolinerakow]

great video. super inspiring! i am using an adaptation of this for an art project, and was wondering how you might go about assigning colors to blobs (say red, yellow, and blue), rather than numbers.

[DamianReloaded]

I think your variable naming convention is facking you up. One word abbreviations starts becoming less and less conspicuous after you get many of them. I personally never abbreviate. The time you save abbreviating variable names multiplied by 100 is the time you lose debugging them later. It was a great video nonetheless, the timeout is a great idea I wanna know how did that turn out. I'm gonna watch the next video... ^_^

[calvinbarajas]

Awesome video, thanks so much for sharing. BTW, I need my computer (Camera, OpenCV, Arduino/Raspberry Pi) to see the shape of a head under both facial and head hair. Can you point me in the right direction please. Can a radio wave sensor capture the contour of the head/face under all the hair? If not, how about infrared, UV, heat? It would need to be pretty accurate.

[federicopanni]

Thanks for the tutorial! I just have a question for you: How can i consider blob 0, blob 1, blob 2 separately? I mean.. I need to say: if blob number 0 goes to a certain part of the screen and blob number 1 in another part of the screen, than println "presence"

Is it possible? Thank you so much, your tutorials are all great!

[cian]

Hi ! I'm wandering which library were you talking about at 0:21 ?

[GameAndGamer]

hi Daniel i have a mac too when i try to execute same code in laptop the video that is displayed is flipped ( i mean text and everything is mirrored )
i tried using this code in draw
pushMatrix();
scale(-1, 1);
image(video, -video.width, 0);
popMatrix();
but the output screen instantly crashes do you have any better method.

[JoostWagensveld]

Hi, Can you also do this in simple C++ without openCV?

[thefrozenindie]

what kind of library does processing use ?

[JohnJ-xmfg]

Daniel Shiffman
Make a version of flappy bird where instead of clapping you flap your hands when holding cups lol