How to Use Webcam In Google Colab for Images and Video (FACE DETECTION)

[@TheAIGuy]

Hope you enjoy! Let me know what other functionality you would like to see in Google Colab.

[@aymankasrawee1431]

Great video about Colab, can you make video for yolov4 API to return image with bounding Box and json details.

[@jothegamechanger]

Thanks maybe next video gonna be good yolov4 on android device sir

[@jesilk.b8920]

That was a amazing video. What are the changes i need to make to connect a colour camera instead of webcam? Can you help me with that

[@ghalibenayad5853]

Face recognition in real time please using the face recognition library or other techniques

[@alloc7260]

Hey bro where did you get that video streaming js code?

[@kunalgupta2960]

I can't wait for the yolov4 version of this. It will really help in my project. You're a lifesaver dude

[@MuhammadWaseem-nx2ln]

Did he uploaded it?

[@maneshwarsingh8688]

Need one for real-time face recognition in Colab too!

[@farjadhaider4401]

Hello. You are such a great helper. I solved a lot of problems by watching your useful videos. Great work!!!

[@TheAIGuy]

Thanks! Appreciate the support.

[@denisusalev]

Thank you for the awesome video, very helpful and well explained! I do however think you should not endorse using the GPU as the 'default' behaviour unless you are doing any real model training, because Google's resources are not limitless, and someone else who might really need the GPU can be getting a limited quota because of that. The rule of thumb I think is that unless you are doing some heavy training, stick to the CPU, live and let live.

[@catafest]

A good idea is to use this with detection of the user and cut the image when another person is in the area of the webcam. I don't have a webcam but I used often colab and is a great tool.

[@JoelBautista]

I can't believe that cuteness and nerdiness can come in one with AI guy. Thanks a lot for this! It helped me with my thesis! I subscribed right away. :) Keep it up :)

[@sayanhazra5946]

Bro that's the thing I was trying to find since last few eeks, u made my day, that was simple but tricky, great man, u are awesome, i just subscribed

[@dieweltentdecker5878]

OMG, that is the craziest video about Colab!

[@TheAIGuy]

Haha thanks! Appreciate the kind words. All the best.

[@dieweltentdecker5878]

@@TheAIGuy Does this method also works with Tensorflow?

[@rodrigobakzan782]

Thanks for make this content, I'm from mexico and I learn about this topic for you. Greetings

[@TheAIGuy]

Thanks for kind words! Glad you enjoy

[@aispectrum]

This was awesome! Easy to follow tutorial and the code ran smoothly like butter.

[@levi-civita1360]

Could you please provide me where you pass camera index, if i want to pass rtsp stream or another camera index . where I should modified the code ?

[@shaunling4800]

Is it possible to do full screen for webcam

[@ashutoshdhanda]

Excelente!! Could the code be used for ip cameras (rtsp/ftp/http) ? if so, where to put the URL?

[@trishalaahalpara918]

Thank you so much for this Video. I am trying to use webcam video for my trained model. And I am getting an error to create the transparent box as there is only two inputs that is getting passed on to the x,y,w,h...I mean thats what the error says. Can you also make a video to use webcam video for our own custom based trained model.

[@viethoalam9958]

dont know why i cant connect to camera - mac air M1 from colab, hope you can give me some suggestion.

[@abdulrahmansabri4547]

Thank you for your efforts. Do you know how to speed up the detection? Thank you in advance

[@yinghaohu8784]

Does this support audio in the same time ?

[@abdelghfarrefat6674]

please whey transparent overlay layer rate is slower than camera frame rate?

[@TheAIGuy]

It will always be one frame behind as you are overlaying on the next frame in the sequence.

[@liamsoccerpro3040]

can you stream the video to a socket to a bad end python server on another device with good fps

[@adamyordan3301]

Hello sir, how can i configure if i have mask detection model and i want use this model with webcam on google colab? thanks sir

[@anupamaranjapalli9999]

in which programming language is used hear either java or python

[@kalyantadala3630]

Could you explain the 68 point facial landmarks detection using the Google colab

[@sweetdiaries4936]

How we can handle mediapipe as we don't have to create a box but a set of landmarks, so how we can update bbox or some other variable to show real-time results with a webcam in colab.

[@harshitrathore1744]

If we have uploaded a video on colab and want to display that with cv2.imshow() how can we do that?

[@FarukhKhalilov]

Thanks for the video tutorial. Is there anyway to split and save live video in small video chunks (< 5 seconds each)? Thank you.

[@Alex-ud4tl]

Great video. I would like to integrate this face detection with the yolo4 deepsort that you have created. How can i detect faces for a video of people walking?

[@ritwiklal2000]

Could you make a video where a functionality to track eye movements can be added alongside face detection?

[@umarsadique7166]

thank you sir good job. I want to do the same process but the camera I am using Hik Vision Camera. I don't about java so I am confused that where to put my camera link. Is it possible to access Hik vision camera in colab?

[@jose3538]

do you know if this works with plain jupyter notebook as well?

[@shubhamvashisth9518]

That's pretty direct, you don't need javascript for that, just use the regular OpenCV methods

[@TheAIGuy]

Yes for normal jupyter notebook you can just run regular OpenCV to open webcam. This is special for Google Colab notebooks.

[@jose3538]

​@@shubhamvashisth9518 even for jupyter server and not run locally?

[@shykhabdull52]

Can it work on Yolo V5?

[@shaunling4800]

@The AI Guy is it possible to have full screen or different window pop out?

[@ashok9588]

How can i change the webcam. I am using an external webcam

[@anilnaidutankala1683]

hello great video but i need to detect hand (for sign language detection) can you please help me

[@vishvambarpanth2888]

Awesome! How to increase the number of frames being collected?

[@hanujirahmatnagoro4518]

excuseme sir... is it possible if we use from youtube video ..? how ..?

[@TheAIGuy]

Not sure i understand what you are asking? Can you try explaining again.

[@hanujirahmatnagoro4518]

@@TheAIGuy : i mean, is tit possible to read/load video from youtube, especially livestream youtuve video n processing it with opencv using google colab..? Fyi: i have a problem when i try to "imshow" the procesesed video in google colab... and the output is something like "generates many picture" from video .. (as we know frame per second).. It will be a great content idea if u want to make the tutorial of it... #soryForMyBadEnglish

[@mumtazprima2959]

hey mr can u build it for yolov3? thanks!

[@atheeralsaif7178]

Greetings, How cv2.VideoCapture(0) works with u? is there any specific prerequisite?

[@yasirahmad5014]

Thanks a lot. Very helpful for me.

[@shubhamvashisth9518]

Can this work with YoloV4 ?

[@TheAIGuy]

My next video will be using this for YoloV4. So yes it can work with Yolov4!

[@Lucas-ri1im]

@@TheAIGuy That's great news! When will this video be uploaded? :)

[@hilalyesilova8624]

@@TheAIGuy when to publish this YOLO version the forecast ? I'm working on a project, and this video is exactly what I need.Thank you for video

[@savasboluk2766]

hello you did a great job, how can i use this with avatarify?

[@TheAIGuy]

I'm not familiar with avatarify? Is it compatible with Google Colab?

[@savasboluk2766]

@@TheAIGuy I have tried. avatarify does not work with google colab. so i did it with another app similar to avatarify, its name is avatars4all

[@AzamAzam-ur8bl]

masking in video by google colab

[@rajpatel1279]

Thank you very much bro 👍👍🥳🥳

[@SR258567]

It is a great video for noobs like myself to learn CV! really appriciate for taking time to make this informative video. while i was running your code in colab, the front webcam of my laptop was working for picture but for webcam the back webcam was activating. i wasnt able to solve this issue. it will be nice of you, if you could guide me on this matter! thank you in advance!

[@oa5818]

You should do face recognation

[@TheAIGuy]

Could definitely add this in in the future. Thanks for the suggestion!

[@helloWorldPlus]

Hello!! I want to be able to identify faces and count them once per each person while on streaming. How can be done? So far I see the model is constantly predicting for the same person and don't identify it as one entity.

[@sheetalprajapat7760]

Hi AI Guy...Hoping you are doing Fine...Can you help me out to connect a trained model(Emotion Detection) with this process....

[@CharcoalProduction]

Can you make a video about homemade LPR camera?

[@poorneshs4222]

Pleases make a video of using webcam with yolo4 in colab ... your videos are awesome pls make this kind of videos

[@damodarang2520]

Thanks a lot. Helpful to me

[@천슬라]

I want to make eye tracking using colab

[@foodhubforkids9767]

Hey! By any chance could somebody help me with the same for an external camera... The code is working but it is showing a black screen instead... I have tried 5 cameras and it is showing the same for every single one of them. If somebody could help me then it would be great! Thanks!

[@abdelghfarrefat6674]

how can flip image likes mirror ??

[@siddharthtyagi1254]

sir, please make one video on ssd

[@TheAIGuy]

Will do! Thanks for the suggestion.

[@mostafaezzat9386]

Can you make a video over action recognition? or video classification using 3D-CNN thank you so much for your videos

[@jangam.raj02]

Thanks mate.

[@001_afshanmassarat9]

thanks a lot ...it was helpful

[@ilmanzidni8847]

Hello Sir, I already train my deep learning model in colab and want to deploy it to the webcam in colab, is there any tutorial to do that task?

[@ahmedm2702]

did you find any solution to this?

[@ilmanzidni8847]

@@ahmedm2702 well actually no, it kept getting error message, i did it by downloaded my model and deploy it on local

[@ahmedm2702]

what version did you use and what was your dataset size

[@allstar2628]

Thank you 😊

[@rajashehryar2002]

Jazak'Allah bro

[@keshav2136]

How to access webcam on DeepNote?

[@saverahanif9139]

my bounding box is not created

[@thecsabel]

Nice video. Unfortunatelly I am struggling with exporting the bounding box coordinates from my own model and therefore I can't use this method. Well expalined though. I might be able to make my code work soon too. Let's hope..

[@UZMAALFATMI]

Thanks!

[@johanverm90]

Well done !!

[@Dr.AhmedQusaySabri]

great video thanks a lot

[@siddhikhanvilkar9890]

Hello Sir, I have learnt use of YOLO by watching your videos. Thank you for providing such videos. Sir can you please provide us video on how to distinguish between static and moving object using YOLO in google colab?

[@vulamnguyen9453]

You have to track those objects first. Yolo is detection model. It detects objects every frames but it doesn't know the relationship between objects in current frame and previous frame. Tracking algorithm will solve this. It will give each object a specific ID. Based on their ID, you can find out which object moving. For example, in first frame, object ID 1 at position (x1,y1). In the next frame, object ID 1 at position (x2,y2). So that the distance it has moved is distance = sqrt( (x2 - x1)**2 + (y2 - y1)**2 ) pixels. When you find this distance, you set a threshold ( for example 10 pixels) if distance > 5 pixels => object ID 1 does move else it doesn't move at all. The threshold you need to try by yourself.

[@태기_개인]

love this! =

[@eng.mohammadshericmrp9251]

Thanks

[@tada_code6370]

thank you

[@HughJackman-w3z]

anyone here to help in 2023

[@Bigboibremmer]

#LEGEND....

[@bnproduction9298]

Why am I left with only 36GB of disk space without even running the cells ?

[@billchen559]

Nice

[@rifqisyekhi]

thanks

[@kadaliakshay6770]

Hi do you want dark theme in google colab

[@雅雁許]

I'like it.

[@c.m.gcontroller7512]

Hello

[@jacsonferneycardenastorres9029]

crack!

[@hitarthpaliwal4341]

Love you man