These would be SO helpful for programming.
Please consider!!!

-Thanks

Well, could you explain a bit on how they'd work?

for vision sensing you could have a block like:

<seeing [ face] ?>

and

go to [ face]

(you could replace face with something like: hand, eyes, mouth, nose, just something like that)

and the same thing for vocal recognition.

These blocks would just sense what you want through the camera.
vision sensing would be used for things like Xbox kinect
and vocal recognition like siri

hope this helps!

panther__ wrote:

for vision sensing you could have a block like:

<seeing [face v] ?::sensing>

and

go to [face v]

(you could replace face with something like: hand, eyes, mouth, nose, just something like that)

and the same thing for vocal recognition.

These blocks would just sense what you want through the camera.
vision sensing would be used for things like Xbox kinect
and vocal recognition like siri

hope this helps!

Ehhhhh… I can't think of enough uses for these blocks: they both seem really, REALLY specific. Also, how would Scratch detect if it's looking at a face, an arm etc.?
Side note: for the Kinect thing, isn't that what the “video motion on this sprite” is for?
double side note: i fixed your blocks

Facial recognition would require pretty complex artificial intelligence…

-ShadowOfTheFuture- wrote:

Facial recognition would require pretty complex artificial intelligence…

…and we haven't even begun to consider privacy concerns.

This requires lots of more technology, including detection and a million scripts to find a single thing.

Also:

cul8er wrote:

-ShadowOfTheFuture- wrote:

Facial recognition would require pretty complex artificial intelligence…

…and we haven't even begun to consider privacy concerns.

You might be recorded…

panther__ wrote:

for vision sensing you could have a block like:

<seeing [ face] ?>

and

go to [ face]

(you could replace face with something like: hand, eyes, mouth, nose, just something like that)

and the same thing for vocal recognition.

These blocks would just sense what you want through the camera.
vision sensing would be used for things like Xbox kinect
and vocal recognition like siri

hope this helps!

I like the Idea!

panther__ wrote:

for vision sensing you could have a block like:

<seeing [ face] ?>

and

go to [ face]

(you could replace face with something like: hand, eyes, mouth, nose, just something like that)

and the same thing for vocal recognition.

These blocks would just sense what you want through the camera.
vision sensing would be used for things like Xbox kinect
and vocal recognition like siri

hope this helps!

https://champierre.github.io/scratch3/ Do it on Google Chrome!
I tried it, it worked!!! Posenet2Scratch Speech2Scratch

pizzaloverkid wrote:

panther__ wrote:

for vision sensing you could have a block like:

<seeing [ face] ?>

and

go to [ face]

(you could replace face with something like: hand, eyes, mouth, nose, just something like that)

and the same thing for vocal recognition.

These blocks would just sense what you want through the camera.
vision sensing would be used for things like Xbox kinect
and vocal recognition like siri

hope this helps!

https://champierre.github.io/scratch3/ Do it on Google Chrome!
I tried it, it worked!!! Posenet2Scratch Speech2Scratch

Yes that is exactly what I want!

panther__ wrote:

pizzaloverkid wrote:

panther__ wrote:

for vision sensing you could have a block like:

<seeing [ face] ?>

and

go to [ face]

(you could replace face with something like: hand, eyes, mouth, nose, just something like that)

and the same thing for vocal recognition.

These blocks would just sense what you want through the camera.
vision sensing would be used for things like Xbox kinect
and vocal recognition like siri

hope this helps!

https://champierre.github.io/scratch3/ Do it on Google Chrome!
I tried it, it worked!!! Posenet2Scratch Speech2Scratch

Yes that is exactly what I want!

DID YOU TRY IT?!

pizzaloverkid wrote:

DID YOU TRY IT?!

Hi, please don't speak in all caps on the forums. It makes it easier to read for everyone.

—

Well, video sensing technically exists. I think vocal recognition wasn't added because Amazon Web Services had it but apparently it couldn't be trusted..?

I don't know, I'm trash with anything related to Amazon.

Nambaseking01 wrote:

pizzaloverkid wrote:

DID YOU TRY IT?!

Hi, please don't speak in all caps on the forums. It makes it easier to read for everyone.

—

Well, video sensing technically exists. I think vocal recognition wasn't added because Amazon Web Services had it but apparently it couldn't be trusted..?

I don't know, I'm trash with anything related to Amazon.

Ok…

I feel like this might not be safe. Combined with cloud variables, scratchers could like, make projects that see people's faces and recreate them and use it to look at people without their permission…

StevenTheSquare wrote:

I feel like this might not be safe. Combined with cloud variables, scratchers could like, make projects that see people's faces and recreate them and use it to look at people without their permission…

I don't think this could happen, because there is no way to sense the color of an object and store that number, so all you would be able to do is see the person's face proportions. (the xy coordinates, not the color and exact shape of the object)

It's not camera detection, but it's a step. I just put together a number detector which detects what number you write down in the box
https://scratch.mit.edu/projects/402223903/

Facial detection isn't very complex, but it is hard to use scratch to detect colors,

so I would suggest using something like canvasjs.com or processing.com

Here's a “simple” facial recognition program I made:
https://scratch.mit.edu/projects/402545703/
change the lock image (top left) and tell it yes when it prompts, then try signing in with the sign in button (top right) try to trick it with other images. Please give feedback on if you were able to trick it

I made a video camera, that can take a video!!!

Before we begin to consider whether or not it's possible, you have to address privacy concerns.Think about how easily you could be recorded using this, no support for that reason and many others.

Edit: Didn't realize how old this topic was. Sorry.