Machine Learning for Facial Recognition in Python in 60 Seconds

In this video we use Mediapipe to run a Machine Learning for Facial Recognition in Python.#python
machinelearning #ai #datascience #python #coding

[@alext5497]

Lib does all the work. We take all the credit.

[@robmulla]

True

[@HB-kl5ik]

I got this video recommended after 10 videos of "How can you earn $5000 per day using chatGPT, chatGPT is craaaziest AI technology and you need to ask it 'How to write a TH-cam Script for..'" I'm not complaining 🙂

[@JimMaz]

No need to reinvent the wheel.
As technology evolves, this will only escalate. You can't be an expert on every subject.

[@HB-kl5ik]

@@JimMaz we live in a world of abstractions for sure

[@anerdtravels]

@@JimMaz Whilst I agree, the more external libraries we rely on the more chance we have for another log4j incident

[@razzakahmad8731]

So many libraries can do this but honestly I always feel more motivated when it is done from scratch like using transfer learning.

[@robmulla]

Yea, fine tuning models can be fun, but you need labeled data which is expensive.

[@JohnWalz97]

Yeah but like that’s not from scratch lol

[@tophat593]

Using a library like that, you may as well open an app on your phone.

[@robmulla]

That's why I love python. 😅

[@Yippie37]

The difference is you can do with this whatever you want next, not just a result

[@alexxander6326]

Please can you make a full video on this

[@KAZZinteractive]

can you expand on this and also teach us how to get mediapipe to recognize facial expressions realtime?

[@zackadam2598]

great short!

[@robmulla]

Thanks!

[@jiyanshsonofdr.rajesh8516]

Great work

[@robmulla]

Thanks!

[@udaykumarrcse2450]

Nice sir

[@robmulla]

Thanks!

[@spll4346]

can this also work on existing footage aswell? instead of webcam footage

[@gabrudude3]

Thanks for sharing. Can the same library be used for full-body 3D mesh as well? For example, if I rotate a camera around 6' tall person and take 120 pictures vs using this library to create that mesh easily?
One more question, what type of camera you used? Do you need LiDAR sensor or just a regular camera to get depth or photogrammetry?

[@robmulla]

I think you are asking about pose estimation? There are a few libraries that can do that like openpose and mediapipe. This is just a normal webcam, no LiDAR!

[@gabrudude3]

Thanks for the response and pointers. I’ll look into those libraries.

[@rqvage4756]

I really love the idea, but I would greatly prefer if I myself made the algorithms. Though I feel like this a good starting point for facial recognition ML.

[@anatoliypotapchuk8174]

Isn't it face detection(find a face on the image), not face recognition?(recognize one face among others)

[@robmulla]

True! Check out my full TH-cam video on deepface for true face recognition.

[@mak2033y]

Great

[@robmulla]

Thanks!

[@gustavojuantorena]

Great! 😃

[@robmulla]

Thanks! 😃

[@dailymemeez]

​@@robmullai am coming to your house.
Check your front porch:)

[@RealHigherLevel]

Rob what software you using in the video to demonstrate how it works?

[@robmulla]

Check out my youtube page for a longer video on deepface.

[@RealHigherLevel]

@@robmulla thanks 👍

[@sivalingeswaranv9159]

Can I get Source code 🤔

[@paul1panter2]

Hi. Can you make a Video how to train a AI so it can identify a face?

[@robmulla]

I have a video about it on my channel! DeepFace

[@goktrenks]

Interesting

[@robmulla]

Thanks!

[@shanksshaki]

How start for beginners ?

[@draculadrac9566]

Can you do a realtime shape detection with codes?

[@robmulla]

You would just need to train the detector on example images of the shapes, but yes it is possible.

[@shafaitahir4728]

but how to save the points then train and load?

[@PhilipDenys]

Wow

[@robmulla]

😏

[@codex1758]

could i use this for authentication ?

[@robmulla]

Probably not good enough for that.

[@tilakraj3082]

I made a face recognition but it hangs after running for 10-20 seconds.....any ways to make it run faster ?

[@robmulla]

Are you using a GPU? If not you might need to invest in one.

[@abidson690]

​@@robmullais Kaggle GPU enough to solve his problem

[@charlessilva5005]

Pls can I get the source code?

[@momsspaghetti21]

How to use this in game

[@nathansuarez5543]

How to extract the face that has been detected in real-time?

[@ahsenkhan3255]

can you find the solution? Because i want to detect persons from video with their names.

[@udaykumarrcse2450]

I want face Emotion detection for song recommendation

[@robmulla]

Check this out. It has a model for emotion but I can't vouch for how well it works: github.com/serengil/deepface

[@udaykumarrcse2450]

Tq sir

[@yehiaibrahim1381]

Hi, i want to build a device that uses image recognition API to help the blind for my school project and idk how to start can you help me ?

[@robmulla]

That sounds like an interesting project. Check out some of my other videos on the topic, hopefully they will help you get started.

[@bonzogaming4333]

CAN YOU PLS SEND ME THE WEBSITE LINK

[@ismailhosenjames]

I will get face encoding only...
Am I right?
But I can't recognize with this.

[@robmulla]

True. It’s recognizes faces within a camera but not who the face is of.

[@sfincione2000]

60 seconds eh? Riiiiight

[@priyasamanta8805]

Whats the code?

[@TauGeneration]

i wonder if it can be used to identify someone

[@robmulla]

Check out my full video on deepface. It's possible but the open source models are pretty bad at it.

[@TauGeneration]

@@robmulla goddamn it. thanks

[@benjaminbilgehan]

i thought that you are a doctor 😂

[@robmulla]

Haha. Oh yea? Why is that?

[@CodX710]

​@@robmullathe kind of clothes you wear

[@robmulla]

@@CodX710 LOL. Ok

[@animania4074]

Its no running nor showing any error

[@jasonnn12]

Is that face detection only and not recognition?

[@robmulla]

Good point. It’s recognizes a face but doesn’t match with a specific person.

[@ismailhosenjames]

I will get face encoding only...
Am I right?

[@MindWhisperer]

hey bro whats, IDE you use ?

[@sfincione2000]

He's using a Jupiter Notebook to write the code. It allows you to write code, put some docs in between and then run individual blocks or the whole program if you need to.

[@MindWhisperer]

Thanks! ​ @sfincione2000. Im gonna try this.

[@adhi_yustisio]

what IDE do you use?

[@robmulla]

Jupyterlab!

[@junkynioy]

I can finally make my own vtubing software 🤡

[@robmulla]

Maybe!

[@anonymosly]

pls share code

[@robmulla]

github.com/googlesamples/mediapipe/blob/main/examples/face_detector/python/face_detector.ipynb

[@luisramon783]

How much you make a year as a ML

[@robmulla]

a few bucks.

[@Avshack]

Sir please can we get source code

[@anyb5020]

Ban facial recognition!!!!

[@wave8060]

Ban you

[@robmulla]

Band of horses.

[@WWG1-WGA]

Now we can clone you and you wouldn’t know

[@robmulla]

I would know though…. Or would I?

[@WWG1-WGA]

@@robmulla keep in mind birth marks. They don’t copy them .

[@jack-bx1xh]

0:00

[@lee7265]

software name

[@robmulla]

mediapipe is a python package.

[@otkurbeck4795]

Source code plz

[@robmulla]

Try googling mediapipe facemesh that’s where I found the source code.

[@Wineblood]

Please not aliased imports

[@robmulla]

Did I use them here?

[@santhnu]

Wanted to see if I would be able to take up Machine Learning/Python from being a Web Developer all these years....decided not to after reading the code written in this video (no offense to @robmulla)...just not my cup of tea, thank you.

[@jacakopl]

In frikkin 5 sec bunch of code just magically appeared 😂. Bs click bait

[@robmulla]

sorry you feel that way. Just intended to show what's possible. I don't explain everything in 60 seconds!

[@theKevatron]

z

[@Reemo3]

This is copying someone else’s code. Lame

[@robmulla]

That is how software engineering works. Shoulders of giants. Making something interesting is the hard part.