Object Detection in 10 minutes with YOLOv5 & Python!

In this video tutorial you will learn how to use YOLOv5 and python to quickly run object detection on a video stream or file all in 10 minutes. This is a great tutorial for anyone interested in using object detection out of the box with python or as an introduction for anyone interested in training their own object detection model in python.

Timeline:
00:00 Intro
00:45 Install YOLOv5
03:44 Detect Webcam
07:07 Detect Video File

My other videos:

#yolov5 #python #machinelearning #objectdetection

[damnhi]

Hey Rob, thank you SO much for this tutorial!! I had been stuck with image detection for quite sometime as many tutorials which I followed skipped certain parts like setting up environments and stuff. But you, my man, you took us step by step through your tutorials! I truly appreciate it and keep up the good work!! Thank you once again!

[chrisberry]

Things like this always convince me I’m on the right path. Thanks 🙏🏼

[BlinkDrive]

I loved the way that you answered everyone in here. Great content.
Great Channel
I subscribed.

[CHRISAT]

Great video, Im currently learning python and cool videos like these motivate me to keep on learning

[kendrickxy]

Thanks a lot for the tutorial! I've been trying to figure out ML for some time now and this video helped a lot.

[Janmejoyslg]

Thank you so much for the step by step guide!

[mohammadraeez]

Hi Rob!! Great video. But I have few questions.
1. How do we train yolov5 model on a custom dataset? For eg: I have a custom dataset of Amazon products, how can I train the model on this dataset?
2. Also, how can I send the prediction results to the front end? I am a backend engineer and is developing a web app for object detection. I know we can send text data in json format, but how do we send the camera feed along with the prediction results to the front end to display?

[pavlostsoukias]

The God is back! Thanks for sharing all this knowlege with us!

[PashkaPashkinson]

fingers are considered as carrots - that was the most impressive AI humour :)

[DataCraftsman]

I love how the model detected the traffic light in the bonnet at 10:16. Possibly dangerous if it was detecting something upside down in a reflection. I actually didn't know about doing pip install -r requirements.txt. Always learn the most random tricks from other people.

[afaquehussain]

thank you for the video, it is very informative and helpful.

[mohammadfarhannurrahman]

Hi Rob! Thanks in detail explanation about yolo.

[VtojeC]

Interesting, sir.

I need to read more about it, but if you have time to share your knowledge with a fool then I'd liek to ask about:
1) Can we specify "what" we want to detect in the image? Let's say I'd like to detect only people and cars, not cellphones, traffic lights, kites, planes, etc.
2) Is it possible to receive bounding box coordinates to the .txt file like TOP_LEFT_X, TOP_LEFT_Y, WIDTH, HEIGHT for each detected object?

I think it is, but would be cool to have confirmation before the research, thanks!

[IgraineDraco]

Great video Rob! How can I take all the object's name after processing and embed those in the video file as metadata? I have a bunch of videos and photos that I want to run this so that I can find them later by searching through tags

[brianhauk]

Hi Rob. Thank you very much for this tutorial. I'm curious to know why you choose YOLOv5 instead of a later version (isn't 7 now released?) and plan on following your example using the version you suggest.

[Julien-hgjh]

So cool ! Is it easy to custom with other dataset ? As it is already train, how many data we need in this new dataset ?

[dr]

great video! How can I do this with my own data set of images and labels? Or add on to the current dataset

[uzbfopf]

I know only yolo is image detection, from your video i got impressed i really want to do but i dont know what interface have u done this and the codes, can u please expalin this step by step 🥺

[alejopaullier]

Hi Rob! Nice video. Is there a way to optimize GPU VRAM utilization rather than modifying the batch size? In my experience most of object detection projects usually require tons of VRAM, even an AWS p2.xlarge seem small for these kind of tasks

[DA_MOZAMAN]

Mr Rob have been training a Yolov4 Model and was wondering if i can use the Dataset i used to train my Yolov4 on Yolov5 or even Yolov7 or do i have to re-annotate all the images into a proper new format ?