Elevate Your Point Cloud Game with State-of-The-Art Depth Neural Networks in Open3D and OpenCV

[@NicolaiAI]

Join My AI Career Program
www.nicolai-nielsen.com/aicareer
Enroll in My School and Technical Courses
www.nicos-school.com

[@luizantonioroqueguzzo8526]

Exactly at the time I started studying about monocular depth estimation you started making these videos, thank you very much. It helps me in my research

[@NicolaiAI]

That's great! Thanks for watching

[@ziaurrehman2180]

Hi buddy I am also interested in pink cloud/3d image processing, can you guide me from where I start if you have any material please share with me , thanks in advance

[@teetanrobotics5363]

Amazing tutorial bro !!!

[@lavanyasureshkannan123]

Is there any way that you can visualize the point clouds it in real time?

[@mbdesign8118]

Where did you get the values for the Q matrix? The Q matrix corresponds to a stereo camera setup but a mono camera is used

[@mush_AV]

Hey this has been really useful in helping get my new project started I was wondering how I could pass images generated in stable diffusion to it instead of my realtime camera input?

[@NicolaiAI]

Yeah u can definitely do that. Just swap the input image to the model with ur own image.

[@unitedealsritwik8176]

Sir! can give some ideas or sample tutorials on Right pair image prediction from left image and depthmap from python (stereoscopic conversation via depthmap in python ) .IT will be very helpful for us .Very nice tutorial with clear instructions.

[@NicolaiAI]

Thanks a lot for watching! Will look into it

[@Pokemon405g]

Hi I love your videos! Could you make a video over how to fine tune models like MiDaS?

[@NicolaiAI]

Thanks for watching! What is the reason u want to fine tune it? Don't really think it will be easy. But they already have fine tuned models for indoor and outdoor

[@microgamawave]

You can make a video about gait recognition biometrics in python
recognized you from your walk model ???

[@HarshMahajan-z5g]

can we complete point cloud with SLAM for real time application, like autonomously navigating the rover

[@dpredie]

Can you show us how to convert the midas3 model to onnx? Or general onnx conversion

[@NicolaiAI]

Thanks for watching! I'll definitely make some videos about how to convert the models and deploy them

[@gauravpatel9382]

How can I get a point cloud from depth with using an intrinsic camera matrix with different f_x and different f_y?

[@luizantonioroqueguzzo8526]

Have you made the .ipynb available somewhere for the point cloud visualization?

[@davidschaupp5423]

How can i get the real depth from the output depth image?

[@afridirahman3906]

I am having this error:
AttributeError Traceback (most recent call last)
in ()
1 output_file = 'pointCloudDeepLearning.ply'
2 #Generate point cloud
----> 3 create_output(output_points, output_colors, output_file)
4
5 cap.release()
in create_output(vertices, colors, filename)
1 def create_output(vertices, colors, filename):
----> 2 colors = colors.reshape(-1,3)
3 vertices = np.hstack([vertices.reshape(-1,3),colors])
4
5 ply_header = '''ply
AttributeError: 'list' object has no attribute 'reshape'

[@Anteater23]

How do install Open3d ? It is nowhere to be found in the anaconda navigator :(

[@潜翔沈]

I’m a Chinese student, and I’m having some problems with your code, what’s the version of the torch

[@NicolaiAI]

Welcome to the channel! I'm using the lastest version from the pytorch website

[@潜翔沈]

@@NicolaiAI Is it a real person or does the machine respond automatically? How can I get in touch with you? I have some qions to ask you

[@NicolaiAI]

@@潜翔沈 it's me 100%! I try to answer all comments on the channel as fast as possible. U can become a member of the channel and then u can get into my private discord and get my personal email where u can contact me

[@潜翔沈]

@@NicolaiAI I have sent you an email and look forward to your reply

[@NicolaiAI]

@@潜翔沈 please contact me through the private discord after u have become a member at the "help with project" level

[@潜翔沈]

Do you have an address for the paper

[@harshahir5901]

do we need depth camera? also can we use this for face recognition if yes please explain.

[@NicolaiAI]

u dont really need a depth camera. but u would need to map the relative depths estimated to real world depths. For face recognition u should just check out facenet

[@harshahir5901]

@@NicolaiAI can you make tutorial on it. thanks

[@潜翔沈]

NameError: name 'output_points' is not defined

[@souvikdatta5686]

Hi, I got the same error, how did you solve it?

[@潜翔沈]

有论文地址吗