Free Anti Spoofing/ Liveliness Detector for Face Recognition System Fake VS Real | Computer Vision

Code and Trained Model:

🚀🚀 My Urdu/Hindi AI YouTube Channel 🚀🚀

Premium Courses:
✔️ Computer Vision Web Development:
✔️ Computer Vision Game Development Course:
✔️ Computer Vision with Arduino Course:
✔️ Advanced Drone Programming Course:
✔️ Learn to Build Computer Vision Mobile Apps:
✔️ Jetson Nano Premium Course:

⚙️⚙️⚙️—-My Gear — ⚙️⚙️⚙️

Follow Me:

#ComputerVision
#OpenCV
#CVZone

00:00 Introduction
02:49 Overview
07:08 Installations
08:20 Testing Face Detector
10:21 Testing Yolo
12:27 GPU
15:31 File Creation Test
17:54 Data Collection Overview
21:21 Bigger Bounding Box
31:50 Face Blurriness
38:29 Confidence Value
41:43 Normalization
50:35 Drawing
52:39 Saving Image
01:07:48 Save Label File
01:13:39 Recheck Label File
01:17:45 Collecting Data
01:27:51 Creating Directories
01:33:48 Get Name List
01:38:51 Shuffle Data
01:40:36 Split Data
01:50:21 Copy Files
02:00:48 Training On Local PC
02:11:07 Training On Google Colab
02:20:56 Drawing

[arjunbali]

thanks sir it is so useful in fintech and because of you I got a job as ML engineer

[riddhimaheshwari]

I had been waiting for this project for so long...it's finally there now...thanks a lot !!

[Nerfzisback]

this is not liveness detection, you just used object detection(yolo) to detect the phone you are holding in your hand and you called that object fake. while object(face) is real. it's smart idea that has potential, but also this way you can pass this system with any face.

[PyTubeMaster]

thank you for teach something that is verey hard to find 😘

[habeshamix]

Can you please make an intruder detection system using face recognition?

[kartikeyasaxena]

Hello i got the args.yaml file but i was not getting the best.pt file in the train weight folder, or running the train.py file. what can be the reason to it? also i was shown an error which i dont know whether was relevant or not, and that was
NotImplementedError: Could not run 'torchvision::nms' with arguments from the 'CUDA' backend. This could be because the operator doesn't exist for this backend, or was omitted during the selective/custom build process (if using custom build).

[Andrei_G]

Super! Thank you! I'm waiting this theme a long time. Great lesson!

[rufiromang]

Minute 52:47 very important point. We have to save 2 things: the first one is our face 😅

[jasonstatham]

Thank you so much sir for this beautiful and Innovative project

[ForTheEraOfLove]

🙏I appreciate you for sharing your efforts to teach us!🎉

[muhammadshifa]

What a great project, thanks for making this beautiful project ❤️❤️❤️

[yasirusenath]

sir you're best tech to us. my problem is all .py fills(train.py, splitdata.py) are run, okay but .pt file can't open it say " the file is not display in the text editor because is either binary use and unsuported text and coding " after run my main.py file its run but not detection I mean real or fake. .pt files to want to run pycharm professional version?

[LazarakisSokratis]

Do you have tutorial for face recognition with data training ?

[jesusfernandotorres]

Hello!! It is a great project and you are a good teacher, do you know maybe the reason why my camera is slow while displaying real and fake faces I'm using the best.pt file that I generated using colab

[wandelmiro]

After training, how can I put it to work on my cell phone or tablet?

[shahhussain]

Sir would you please create a same project with JS also.

[RickSlaughter]

Are you using PyCharm Community Edition or PyCharm Professional?

[askelardd]

Perfect, thank u so much, this helps me a lot

[AbdulRehman-qymt]

is your model overfitt..? As it always detect new person as fake. Also can you tell how much epochs do i need if i have total of 5000 real images and 5000 fake images of 30 different people

[MPChannel-clbw]

awesome, this is exactly what i was looking for for my project