Object detection with Python FULL COURSE | Computer vision

Code: github.com/computervisioneng/...
0:00 Intro
0:50 What is object detection
5:03 Object detection metrics
32:13 Train Yolov8 on custom data
1:36:25 Train Detectron2 on custom data
2:23:56 Object detection with AWS Rekognition
2:57:08 Project 1: Automatic number plate recognition with Yolov8
4:07:40 Project 2: Object detection web application with Python, Streamlit and Detectron2
4:35:01 Outro
#python #computervision #objectdetection #streamlit #detectron2 #yolov8 #awsrekognition #objectdetectioncourse

[@ComputerVisionEngineer]

Did you enjoy this video? Try my premium courses! 😃🙌😊 ● End-To-End Computer Vision: Build and Deploy a Video Summarization API bit.ly/3tyQX0M ● Hands-On Computer Vision in the Cloud: Building an AWS-based Real Time Number Plate Recognition System bit.ly/3RXrE1Y ● Machine Learning Entrepreneur: How to start your entrepreneurial journey as a freelancer and content creator bit.ly/4bFLeaC All my premium courses are available to the Computer Vision Experts in my Patreon. 😉 www.patreon.com/ComputerVisionEngineer

[@user-yh9bl9wn2v]

Do you explain in Patreon the code you used to detect the alpacas on the short videos you used? I coudn't find it

[@ComputerVisionEngineer]

@@user-yh9bl9wn2v hello, do you mean the tutorial about yolov8? I didn't create patreon content explaining the code I used to detect alpacas on the short videos I used. The script iterates in all the frames of the video, it detects alpacas and draws text on top of the frame. Do you need help with that script? Please ask for assistance in our discord community. 😃🙌

[@LuizFelipe-jz2dn]

Thanks friend, because of you I got my 1st job in computer vision in healthcare.

[@ComputerVisionEngineer]

😃 That is so cool! I am so happy my videos helped you getting your first job in computer vision in healthcare! 🥳🙌

[@duppy404]

Congrats!

[@TheBrn12]

Hi! What did you study? Did you just learn online about cv and then applied to the job? I want to get into this field

[@sreekartammana]

That's awesome, congratulations and all the best bro🎉

[@LuizFelipe-jz2dn]

@@duppy404 thanks!

[@hamzawi2752]

A million thanks, that is kind of you to dedicate your time to recording these videos. Merci from Grenoble, France :)

[@ComputerVisionEngineer]

😃🇫🇷 My pleasure! Glad you enjoyed it! 🙌

[@Pablo-qe7zm]

Gracias por el tiempo que inviertes en crear contenido de gran valor y explicado de manera sencilla y clara, soy fan!

[@ComputerVisionEngineer]

Es un placer! Me divierto mucho haciendo los videos! Gracias por tu apoyo! 😃🙌

[@AllanMcKay]

Thanks for the indepth lessons, love it! Keep up the great work

[@ComputerVisionEngineer]

Thank you for your support! 😃🙌

[@antoniopiotufo2247]

Amazing video! All the concepts are explained clearly! I would like view a video when you show how to train a model from zero (not pre-trained model) and take it to best performance!

[@ComputerVisionEngineer]

Thank you!! Great suggestion! I will keep it in mind for future videos. 😃🙌

[@mpfmax0]

muchas gracias, con tus videos logre completar el proyecto de mi curso de intro a machine learning

[@ComputerVisionEngineer]

De nada! Me alegra saber que los videos te han ayudado! 😃

[@keepcool8879]

suscribed immediately, thanks for this !

[@itaydagan7459]

Amazing channel, thank you!

[@shamaldesilva9533]

Wow this is incredible , thank you so much 🎉🎉

[@ComputerVisionEngineer]

You are welcome! 😃🙌

[@bfouad]

Many thanks good man, actually i was going to make free space detector on any wall so, Could you suggest what is the best model I mean Yolo or AWS ?

[@danilzubarev2952]

Thank you, Felipe. Much love :3

[@ComputerVisionEngineer]

You are welcome!! 🥰🙌

[@microcontroller_garage5387]

Thanks bro, your videos helped me in my BTech project

[@ComputerVisionEngineer]

That is awesome! Glad the videos are helpful! 😃

[@MIGUEL-bv1qh]

Well done my friend. It will be nice a theorical video about object detection like to detect multiple labels in a specific object (Ex. Two person in a motorcicle)

[@ComputerVisionEngineer]

Thank you for your support! Ok, noted, a theoretical video to detect multiple labels in an object, I will try to do it. 🙌

[@timowidyanvolta]

Thank you man! This is means so much to me

[@ComputerVisionEngineer]

😊 I am happy you found it helpful! 🙌

[@xuzhipeng6764]

In this video, you demonstrated downloading images and then annotating them. Is there any other way to directly download annotated datasets, such as alpacas.

[@gamerlight27]

man, i just love u i tried to find tutorial in my language(im from russia) and i failed, so i decided to find english video and here you are, huge thank you!

[@ComputerVisionEngineer]

Glad you enjoyed it! 😃🙌

[@gamerlight27]

​@@ComputerVisionEngineer oh and 1 question, can i add new type of object without total retraining?

[@trinadhinkulu3371]

Model A performs better than Model B.since both IOU & mAP are higher

[@ComputerVisionEngineer]

That is correct! 😃🏆🙌

[@titangamer5157]

Thanku sir🙃🙃

[@justj-in5yq]

Hello Sir how to build custom model which is using in number plate detection can you provide detail videos and code for model i did understand whole think but some blocked in model please give some guide !!

[@joelbhaskarnadar7391]

Interesting Work

[@ComputerVisionEngineer]

Thank you! 🙂🙌

[@xxBLAXTERxx]

Muchas gracias! muy bien explicado. Tienes algún tutorial de cómo hacer transfer learning y fine tuning de modelos pre entrenados con Yolov8?

[@ComputerVisionEngineer]

Gracias! No tengo videos sobre como hacer transfer learning de modelos entrenados con yolov8. Tal vez lo haga para algun proximo video. 🙌

[@sumitsamanta7416]

can u provide with the dataset u used ? plzz

[@chukwuntachibueze4281]

Great video boss. However, I am having issues generating the dataset from github. Are there documents or folders that need to be created before running the codes after downloading the attached CSV files? Because, the second code on "create dataset yolo format" runs an error: FileNotFoundError: [WinError 3] The system cannot find the path specified: '.\\data\\train. Kindly help me decipher the issue. Thank you.

[@ComputerVisionEngineer]

Hi, thank you! Do you mean how to download an object detection dataset from the google images dataset v7? I created an additional resource, which is comprised of a video tutorial with step by step instructions and also a Python script which makes the download process super straightforward. The resource is available to my Patreons here: www.patreon.com/posts/how-to-download-91285241

[@lucastyrrell5344]

Hi great video would not be able to make something without it, however I ran into an issue when training the model where it would not create a model directory with a .pt file meaning that it could not detect anything. If you have any advice or know what could help it would be much appreciated.

[@ComputerVisionEngineer]

Did you have any error during the training process?

[@lucastyrrell5344]

Hi managed to fix the issue just used a larger dataset with more epochs and it now works. But not sure what I would’ve done if I didn’t find your video, so thank you!

[@kamrangul2876]

I am just here to mark my attendance. Thank you sir 😘

[@ComputerVisionEngineer]

😃 Thank you for attending the course!! 💪🙌

[@dongivafoc4493]

Hi could this be used to build an attendance system using facial recognition with multiple classes which would be the different people ?

[@ComputerVisionEngineer]

Hi, take a look at the face_recognition Python library, it is going to help you a lot with the face recognition. 🙌

[@thetechmachine5446]

where can i get the dataset?

[@buni.s]

are those pycharm features available in community edition??or I'd need to install professional IDE??

[@ComputerVisionEngineer]

I am using the community edition

[@buni.s]

@@ComputerVisionEngineer oh thanks

[@varungoel7326]

Hi! Can we pass bounding box coordinates as input to our model received from another model as input (either as a frame or image?) in yolo v8 It will be of great help if you could tell….

[@ComputerVisionEngineer]

Hey, the input of an object detector is usually an image, given by either the image path, or a numpy array or similar object. What do you mean by passing a bounding box as input?

[@varungoel7326]

@@ComputerVisionEngineer I am getting the bounding box coordinates of the detected vehicles using my model. Further using these detections(Bounding box coordinates) I want to get these vehicle's number plates respectively using another model using yolov8. So is there any way to pass these bounding boxes as input to the number plate detector model without saving the detections into images...??? Thanks alot for replying to the previous comment...!!!

[@ComputerVisionEngineer]

You could crop the frame with those bounding box coordinates so you keep only the vehicle, then input the frame crop into your additional model.

[@varungoel7326]

Okay,thanks alot for your help And thanks alot for replying…

[@yuashinsewsunker4726]

hii , want to know if i should get started with YOLOv8 or openCV

[@ComputerVisionEngineer]

Both yolov8 and OpenCV are good technologies to get started in computer vision. 🙌

[@PANDURANG99]

is there face recognition?

[@HarshSingh-zp7jb]

Hey, how can I get the exact same dataset used in this video for YoloV8 tutorial

[@ComputerVisionEngineer]

If not mistaken the data is in the github repository, if it is not send me a message through discord to remind me I upload the data.

[@Thenvi11]

Does anyone know how to add new images to an existing task? @ComputerVisionEngineer

[@nyamdavaanmd8708]

Really appreciate the video. Is it possible that you share the code which recognized alpaca from videos? It was mentioned at 1:32:52

[@ComputerVisionEngineer]

If not mistaken the code is in the github repository.

[@judeanpaulyn]

@@ComputerVisionEngineer I checked the repository. Unfortunately, its not there. Requesting the codes you used for recognizing alpacas on videos please, Sir.

[@tharealdonzw]

Thanks from Africa Zw

[@ComputerVisionEngineer]

You are welcome! 😃🌍🙌

[@mohamedmedo5476]

how to download dataset please

[@hamzayousaf5538]

Dataset '/Users/Hamza Yousaf/teeth detector/config.yaml' error object of type 'NoneType' has no len() kindly help me i face this error

[@ComputerVisionEngineer]

Not sure what could be going on, is this about yolov8 object detection? you could create an issue in the repository of that project. github.com/computervisioneng/train-yolov8-custom-dataset-step-by-step-guide

[@hamzasalhi5362]

Hello professor, could you create a project dedicated to detecting damages in a car?

[@ComputerVisionEngineer]

Hey, that it's a very good idea for a future video! I will try to. 😃🙌

[@rashedkord6697]

Hi plz make a video about sam model.

[@ComputerVisionEngineer]

I will try to 🙌

[@HarshSingh-zp7jb]

24:27 Model A will be better since it has a better mAP and IOU score

[@user-om9qg7yj9m]

In the section with YOLO, you did not show anything because how to apply a trained model, nothing because where the trained model lies.

[@ComputerVisionEngineer]

Hi, do you mean you would like an example regarding how to make inferences with yolov8 object detection?

[@storage50kb]

I have some ideas for same project to make it more useful. If you interested let me know

[@xuzhipeng6764]

How to Download Marked Datasets

[@ComputerVisionEngineer]

Hey, do you mean the datasets I use in the different tutorials in this video? The datasets are usually in the github repository.

[@chukwuntachibueze4281]

@@ComputerVisionEngineer Yeah, but the process you stipulated on how to download it is quite confusing.

[@furkandemirel9595]

model aa

[@istiakahamedtushar7094]

Which model performs better? Ans: Model A.

[@Intellectualmind4]

Ola amego

[@ComputerVisionEngineer]

Hola amigo! 😃

[@cheapearth6262]

💀💀 his accent is awesome

[@ComputerVisionEngineer]

😄 Glad you enjoyed it! 🙌

[@koushikroy2853]

Obviously, A is better than B

[@ComputerVisionEngineer]

That is correct! 😃

[@gkhnmr]

please stop saying this 'veryveryveryvery' in every simple sentence :D Please don't be angry with me. I really liked your video. Just an advice. :)

[@ComputerVisionEngineer]

Lol, thank you for your feedback. 😃🙌

[@buni.s]

Model A is better