Відео

Pued usar esto con three js?

I saw the term Gaussian Splatt and looked it up. I came to this video and watched a few minutes of it. I understand that this chap is speaking English - I recognise some of the words - but honestly, on the whole it sounds like a sequence of random words "git, splat, hub, fork, repo". Great stuff!

Has anyone encountered this error? splatting>python train.py -s images/input Optimizing Output folder: ./output/be20adee-d [13/11 10:03:31] Tensorboard not available: not logging progress [13/11 10:03:31] Traceback (most recent call last): File "train.py", line 282, in <module> training(lp.extract(args), op.extract(args), pp.extract(args), args.test_iterations, args.save_iterations, args.checkpoint_iterations, args.start_checkpoint, args.debug_from) File "train.py", line 51, in training scene = Scene(dataset, gaussians) File "D:\Library\gaussian-splatting\scene\__init__.py", line 49, in __init__ assert False, "Could not recognize scene type!" AssertionError: Could not recognize scene type!

Nice display, thank you.

I really enjoy your videos! Keep up the good work my friend.

For those who have problems with the right CUDA version, my advice is downgrade your Visual Studio to 2019, since NOT all versions of the visual studio 2022 have the supported MSC_VER for CUDA 12.x or below

awesome and thanks

can you make a tutorial about making dynamic gaussian splattung?

I get this while "Installing pip dependencies" Pip subprocess error: ERROR: Directory 'submodules/diff-gaussian-rasterization' is not installable. Neither 'setup.py' nor 'pyproject.toml' found. failed CondaEnvException: Pip failed

There is so much depth information with the parallax effect and lighting/shadows.

Do you also have sometime the pinhole bug with GLOMAP when trying to train Gaussians splats?

Hi, thank you for your awesome work. I want to know if you ever try to use 360 video of indoor environment for Gaussian Splatting and is it ok in term of output quality?

Hi, I have a problem during Installing the Optimizer. When I run the second command ((conda env create --file environment.yml) it starts ok but then gives error while installing pip dependencies like below: (any help is appreciated) Pip subprocess output: Processing c:\users\tarek\gaussian-splatting\submodules\diff-gaussian-rasterization Preparing metadata (setup.py): started Preparing metadata (setup.py): finished with status 'done' Processing c:\users\tarek\gaussian-splatting\submodules\simple-knn Preparing metadata (setup.py): started Preparing metadata (setup.py): finished with status 'done' Building wheels for collected packages: diff_gaussian_rasterization, simple_knn Building wheel for diff_gaussian_rasterization (setup.py): started Building wheel for diff_gaussian_rasterization (setup.py): finished with status 'error' Running setup.py clean for diff_gaussian_rasterization Building wheel for simple_knn (setup.py): started Building wheel for simple_knn (setup.py): finished with status 'error' Running setup.py clean for simple_knn Failed to build diff_gaussian_rasterization simple_knn Installing collected packages: simple_knn, diff_gaussian_rasterization Running setup.py install for simple_knn: started Running setup.py install for simple_knn: finished with status 'error' Pip subprocess error: error: subprocess-exited-with-error × python setup.py bdist_wheel did not run successfully. │ exit code: 1 ╰─> [22 lines of output]

Thanks for the video. One quick question though, why is it that I can see gaussian splat in view port but not in the main camera. In other words, Unity does not render gaussian splat in any of the camera except view port. If you know, please share. Thanks

Heya, I've followed your tutorial until I've had to pass the command convert.py -s data/Videos/input The directory exists, I'm just not sure what "Invalid options" that I've provided that caused the python command to fail. E20240928 01:38:44.370547 19964 logging.cc:56] [option_manager.cc:811] Check failed: ExistsDir(*image_path) E20240928 01:38:44.371256 19964 option_manager.cc:877] Invalid options provided. ERROR:root:Feature extraction failed with code 1. Exiting.

Heyya! Thanks for providing such a great tutorial video. I wonder if you have no video for MacOS?

Dear Jonathan. I have a question. When you cut the pano images with meshroom into cube maps they will have the size 1200by1200 with your line of code. I wonder if there is a formula with which one can calculate the maximum size possible for the input pano. I for example will be able to use a 8k 360 camera soon and i wonder what would be the ideal cube map size for the corresponding input material. Do you have any idea how to calculate or figure this out? Or is it a simple try out process? Thanks :-)

Building wheel for simple_knn (setup.py): started Building wheel for simple_knn (setup.py): finished with status 'error' Running setup.py clean for simple_knn Failed to build diff_gaussian_rasterization simple_knn Installing collected packages: simple_knn, diff_gaussian_rasterization Running setup.py install for simple_knn: started Running setup.py install for simple_knn: finished with status 'error' Pip subprocess error: error: subprocess-exited-with-error plz help

Building wheel for simple_knn (setup.py): started Building wheel for simple_knn (setup.py): finished with status 'error' Running setup.py clean for simple_knn Failed to build diff_gaussian_rasterization simple_knn Installing collected packages: simple_knn, diff_gaussian_rasterization Running setup.py install for simple_knn: started Running setup.py install for simple_knn: finished with status 'error' Pip subprocess error: error: subprocess-exited-with-error plz help

Hello there, I recently followed your instructions, but I encountered an issue with Nerfstudio. It requires a YML file using the "--load-config" option, but as you know, the 3DGS process doesn't produce a YML file. As a result, it's difficult to follow along with your video now. Is there any way to open the 3DGS PLY file using Nerfstudio? I'm quite new to this field, and thank you for your great video regardless!

Can this be used with different feature extractors?

is it possible for you to share the 360 video for practice. I haven't been able to find good 360 videos to try gaussian splatting on. I have tried it successfuly on a lot 2d videos but just can't seem to find a good 360 one. Thanks for you the beginner guide, it was really helpful

Thank you!well ,could I use it to get a "transform_json" file? For example, for instant-ngp training

does it support dynamic scenes ?

nice. can anything that gets produced via "dark magic" you've showed be used with Post Shot ?

Building wheel for diff_gaussian_rasterization (setup.py): finished with status 'error' --> Always this error is coming while running the --> conda env create --file environment.yml

Special thanks from Japan!!!!! Arigato

Thank you so much for sharing this stuff. Very powerful! I wish you would make a video describing how to get great quality gaussian splatting for a room. Tips like what kind of videos or what kind of pictures. "How to take good quality source material for great splatts" or something like that. Thanks!

hey Jonathan, great tutorial! I am having trouble when trying to run this with nerfstudio, I run the script you have and it works great and then I ran the same ns-process-data command. However, when I started training with ns-train, all the cameras were positioned in the same spot. Checking the transforms.json file showed that all these cameras were positioned at the same spot. Is there some way to fix this? wish you could show that training with nerfstudio part in your video

@thenerfguru You ran environment.yml directly which installed submodules/diff-gaussian-rasterization and submodules/simple-knn I am facing issues in installing this setup I have ran nvcc--version cmd which gave cuda 11.8 tool kit is installed please do support on what changes I need to do I tried it multiple times gone through multiple opensource communities anyone who has done successfully can answer

To me the way glomap works is really similar how 3d scanners align scanned frames, or rather correct them after scanning with a global registration ICP algorithm. As shown in the video this will easily work with objects with distinct features, but if you tried it with geometrical shapes or featureless/smooth objects it most likely wouldn't. Still very interesting! Didn't think someone was insane enough to try matching "stereo frames" for photogrammetry but for GS it makes sense I guess as you don't need the mm accuracy. Or perhaps I'm completely wrong, gotta read that paper 😂

Hello, do you think it would be possible to create a complete race track and then map it using Blender etc.?

it took a lot of time and i kept doing the same steps over and over because you didn't speak in a clear way so some times i didn't understand you in general I'm not programmer or anything of that and i didn't install visual studio before but in the same time i didn't watch any one else because i loved your way i don't know why keep going you are amazing honey

This is really interesting. I'm excited to explore this! By the way, great job, @thenerfguru. I hope you create videos on dynamic scenes as well.😅

(( CondaEnvException: Pip failed )) keeps saying that every time nd idk what is the prplm >_<

You Rock Jonathan 💪🏽

Great tutorial Jonathan! Glomap seem to be very fast. Great work with the python script to make this even faster. Thanks for that. I have to check how this could be implemented with Postshot.

Awesome video. 15:12 @anyone trying this: Be sure to not have whitespaces in your file path as it will not work (as of now)

I'm grabbing 🍿 for this one... Thanks !!

"Hi there, great video! I'm currently using a reality capture pipeline to generate point clouds. I'm curious to know more about how this method compares to traditional reality capture. Could you elaborate on the differences in terms of processing speed and quality of the resulting point cloud?

you dropped this, 👑

can you export this into unreal engine ?

C:\Users\Administrator>conda create --name nerfstudio -y python=3.8 'conda' 不是内部或外部命令，也不是可运行的程序 或批处理文件。 C:\Users\Administrator>

Hello there. It's a great video on how to use it, especially with installation and this one. Can we export a mesh from Gaussian splitting? There may be an option in the viewer under the point cloud. I haven't used it yet, so I reviewed your tutorial. Thanks.

What happens if i run it on a gtx 1650???? will my pc explode??

I was here before it went mainstream. It's awesome.

Thanks for your tutorial!! Could 3D Gaussian splatting interact with light in the scene, and could it be captured by a reflection probe?

You need a tutorial building everything from source to run on Linux Distro.

Amazing effect for static backgrounds

During the process of using gaussian-splatting-Windows, when the image files originally in the 'input' folder are processed by python convert.py -s and output to the 'images' directory, they reduce in number from dozens to just a few, and these images now display a wide-angle effect. What could be the issue, and how should one go about troubleshooting this?