Deep Drowsiness Detection using YOLO, Pytorch and Python

Want to leverage YOLO object detection for safety?

One great implementation is using it to determine when drivers might be feeling a little drowsy. In this video we’re going to do exactly that using a fine tuned, customer object detection model powered by YOLO and PyTorch!

In this video you'll learn how to:
1. Instal Ultralytics YOLOv5
2. Detect Objects from Images
3. Detect Objects from Pre-Recorded Videos
4. Detect Objects in Real Time Using OpenCV
5. Fine Tuning a Drowsiness Model using YOLOv5 and PyTorch
6. Perform Real Time Drowsiness Detection

Get the code:

Links

Chapters
0:00 - Start
0:48 - Introduction
1:18 - Gameplan
2:23 - How it Works
3:05 - Tutorial Start
4:12 - 1. Install and Import Dependencies
10:51 - 2. Load Model
13:44 - 3. Make Detections using Images
21:05 - 4. Real Time Detections and Object Detection using Videos
30:05 - 5. Train a Custom YOLO Model
1:10:28 - 6. Detecting Drowsiness
1:17:58 - Ending

Oh, and don't forget to connect with me!

Happy coding!
Nick

P.s. Let me know how you go and drop a comment if you need a hand!

[superchargedhelium]

This is one of the best tutorials I have followed. Not only for drowsiness detection but also just the fundamentals of data capture, labeling, custom model training. Thanks!

[abimanyuvijay]

Hands down the best yolo tutorial there is out there. Extremely clear, straightforward and fun to watch. Really helped me out with my internship project.
It would've been even better if you had included some sort of a warning beep when it detects you to be drowsy.

[muhammadbustomi]

I'm from Indonesia and really like all your content. All of your content is learning that can't be paid for cheaply. Thank you sir.👏

[kuark]

Whenever i need to learn something new about ML, i always come to this channel. I rarely comment to videos but this guy is just awesome.

[mikeb]

Recommended. Nice video all round. Pitched at the right level, clearly explained and kept it interesting. Nick goes through intro of using the model and then the framework of custom training and tests on that trained model. Thank you.

[ethelyulianatejadarojas]

It was your tutorial that allow me to do my first experiments with YoloV5, I liked the clarity and the transparency, many other tutorials kind of hide things but you were coding, correcting/ adjusting your code and showing your environment which is priceless. Thank you

[satoshinakamoto]

This looks way more convenient and easy to implement compared to tensorflow using protocol buffers.

[josephenchill]

You saved me again. This channel is underrated!!!

[mohammadghanatian]

Thanks for your great video. I have a comment. At 25:11 when you used the webcam video, the images were in BGR format (you used opencv) and you didn't convert them to RGB before giving them to the model. Considering the fact that the model was trained on RGB images, by converting the images to RGB we may get better results.

[reubenthomas]

This was soo epic! Please upload more! I was glued to my laptop as I tried this out. You're a great teacher! :D

[grzegorzkozinski]

That's exactly what I was looking for, for past few weeks. Thanks!

[shaikabrarulhaq]

A gentle remainder !
The first coommand to install Pytorch is about 3GB+ so wait for some time...

[henkhbit]

I am always awake when i watch your video Nick. Nice example with yolo!😃

[adityanjsg]

Thanks is too small a word!! With this tutorial I solved a major industry requirement.

[pragyamishra]

What an amazing tutorial I came across. Explained in such an efficient way. Thank you.

[satoshinakamoto]

Wow, so many amazing projects! Thank you Nicholas.

[ravind_]

This is one of the best tutorials I have followed.

[YoungZ]

Thanks for the excellent tutorial. I followed all your steps in the video and got a great result!
I already left my like!

[imdadood]

Good to see Michael Scofield is into Deep Learning after breaking out from the prison.

P.s. I love your contents!

[einstein-munachi]

I'm about to work with this new knowledge to build my MSc. Data Science & Analytics thesis project.... Thanks!