Real-time Instance Segmentation on Webcam with PixelLib and PointRend

You will also get access to all the technical courses inside the program, also the ones I plan to make in the future! Check out the technical courses below 👇

_____________________________________________________________

In this video 📝 we are going to talk about Real-time Instance Segmentation on Webcam with PixelLib and PointRend. First of all, we are going to see how to install PixelLib and then we will segment images from a webcam. Throughout the video, we are going to look at the results and also do instance segmentation of multiple classes.

If you enjoyed this video, be sure to press the 👍 button so that I know what content you guys like to see.

_____________________________________________________________

_____________________________________________________________

_____________________________________________________________

📞 Connect with Me:

_____________________________________________________________

🎮 My Gear (Affiliate links):
🖥️ Desktop PC:

_____________________________________________________________

Timestamps:
0:00 Installation
1:53 Instance Segmentation
6:43 Real-time Instance Segmentation

Tags:
#instancesegmentation #semanticsegmentation #segmentation #ai #computervision

[NicolaiAI]

Join My AI Career Program
Enroll in My School and Technical Courses

[mikelgorostiza]

Hi, what is the hardware you're working on in the video? I have to estimate the process time for an specific hardware and this info would help me, thanks.

[sanjaiivijayakumar]

Hey Nicolas, I have been trying to do segmentation in real time with lot of different models and every model compromises the frame rate. YOLOv5 was great but I was not able to access the mask coordinates with which I have to do further processing. Does pixellib provide with the mask coordinates?

[randalllionelkharkrang]

Thanks for this. Im actually looking for face segmentation models that is able to segment nose, ears lips etc. Is there something of that sort?

[tonibruda]

Hey Nicolai, I have question. Im working on my final project for college and need help with code. I have robot and cam pose (30 poses - base2gripper and cam2target) and I need to calibrate camera. Don't know how to make it. What tier I should get on Pateron so you can help me?

[niklasschmitt]

Love your videos! Is it possible to use cvat also for data labelling

[iclick]

Nice. I'd like to see how to add labels (e.g. detect button, link on a computer screen) in the simplest way e.g. with minimal-to-zero fine-tuning/training

Also, in creating new labels, if it requires manual labelling, show us the easiest way to label an instance-detected video like output.mp4 here

[nazirjama]

Hey thanks to you for the good content. I've been following your content for a while now and it helped me a lot thanks.
if came to this lecture basically i have followed all the steps you followed which are not complicated but this is the error that i an getting AssertionError saying 'AssertionError: Checkpoint pointrend_resnet50.pkl not found!).
Is there anyone who is getting the same error as mine? . Nicolai can you help.

[ianlaw]

Hi Nicolas,
Thank you for your video. I tried doing exactly what you show on the video on GoogleColab, but for the line
ins.load_model("pointrend_resnet50.pkl", confidence = 0.7, detection_speed = 'fast')
i got the an AssertionError saying 'AssertionError: Checkpoint pointrend_resnet50.pkl not found!'
I looked up on Github and tried !pip3 install fvcore==0.1.1.dev200512
but still doesnt solve the problem.
Do you have any idea how to solve it ?