You are really a Diamond 💎 for Advanced Webdevlopment I am very lucky to found you sir ❤ I am using Your ideas and Secret Models in my project is looking like future 😂 generation website.
really fun stuff! i used to follow coding train videos really closely years back, fantastic to see you've still got it haha :) i was actually reading through a random old blog post on web archive, and it linked to the processing page from back in like 2017, and i saw you in the thumbnail on the front page - gave me the idea to come back to see how things were goin' & glad i did!
For the next one, would be cool if the drawn dots change in size with the confidence (larger dots to cover more area when less confident) to see it a bit more visually I suppose this only makes sense if the points have their own confidence though lol
Great Job Dan. This is great technology and could potentially make the mouse and keyboard a thing of the past. Have the computer operating system read your hand gestures to control the computer maybe?
In far galaxy long ago I wrote old computer vision algorithms in opencv before this AI boom and everything works fine. Interesting which method is less cpu/gpu intensive and precise....
I am asking to community and it's a general question. Let's say I have three objects and I need to track their positions relative to each other. Should it be an array property in object class like neighbourProximities oorrr a completely seperate array? Keeping a seperate array looks more effcient to me in terms of memory because I'll have only 3 values in array. On the other hand I'll have 2 values in each object which will result in 2×3. And most of them going to be the same. I decided to use seperate array thanks for me for the answer 😂
I tried to write an OBS plugin or a userspace v4l app using ml5.js, but it really feels like they want some browser only DRM... but also they want source available, it's really confusing. How can I use these models in my x86_64 rust application/libraries?
You are really a Diamond 💎 for Advanced Webdevlopment I am very lucky to found you sir ❤ I am using Your ideas and Secret Models in my project is looking like future 😂 generation website.
as soon as the video started, I immediately set hand.thumb_tip y value to the highest possible. awesome coding and explaining as always, sir!
really fun stuff!
i used to follow coding train videos really closely years back, fantastic to see you've still got it haha :)
i was actually reading through a random old blog post on web archive, and it linked to the processing page from back in like 2017, and i saw you in the thumbnail on the front page - gave me the idea to come back to see how things were goin' & glad i did!
It was interesting as always. My best part is where you introduce how this could be used in various ways.
I love, how unbelievable good video tracking got. Even with simple solutions like this, it's perfectly fine for most use cases.
Once we could start using ML, it all jumped ahead a bunch.
As always well explained and easy to follow. Loved it.
Awesome and incredible!
Never clicked a video faster. Pd: I'm really enjoying the nature of code book!!! Thank you so much
Oh hey, its the Kinect!
Surprise Dan video for an otherwise dull Tuesday!!
I'm loving this! What a great t8me to be alive 🎉
For the next one, would be cool if the drawn dots change in size with the confidence (larger dots to cover more area when less confident) to see it a bit more visually
I suppose this only makes sense if the points have their own confidence though lol
Incredible!
Great Job Dan. This is great technology and could potentially make the mouse and keyboard a thing of the past. Have the computer operating system read your hand gestures to control the computer maybe?
Nice
Wow your camera image is fantastic and there's no latency or stutter as it overlays the pose. Do you have lots of lighting or what?
In far galaxy long ago I wrote old computer vision algorithms in opencv before this AI boom and everything works fine. Interesting which method is less cpu/gpu intensive and precise....
Me tooo!!!!
I am asking to community and it's a general question. Let's say I have three objects and I need to track their positions relative to each other. Should it be an array property in object class like neighbourProximities oorrr a completely seperate array? Keeping a seperate array looks more effcient to me in terms of memory because I'll have only 3 values in array. On the other hand I'll have 2 values in each object which will result in 2×3. And most of them going to be the same. I decided to use seperate array thanks for me for the answer 😂
I did a lot of prototype for this before AI was cool
Can we track body AND hands at the same time?
@@richarddecosta that should be possible, just create variables to hold the body and hand tracker.
I tried to write an OBS plugin or a userspace v4l app using ml5.js, but it really feels like they want some browser only DRM... but also they want source available, it's really confusing. How can I use these models in my x86_64 rust application/libraries?
nit: for (const someItem of someArray) does not need a preceding length check, it just won't do anything if length === 0
I gotta do the emoji one, pls ppl remember me this :)
You are so handsome 😊
Sign language translator
Was thinking the same. Could convert into AI voice for those that can't sign.
@BeenThrDunThat you even thought one step ahead of me with the voice thingy... I was only thinking sign to text.
@@jaobillionaire honestly that alone would be valuable. If you have the ambition I'd say go for it!
[0,0,1,0,0] 🤪