01/17/2025
FROM:
=====================
PYTORCH
Conv2d and ConvTranspose2d explained: conv_arithmetic/README.md at master · vdumoulin/conv_arithmetic · GitHub
Convolutions Math 1603.07285
3DGS
VistaDream
CUDA for PyTorch
Pytorch GPU Setup Guide | The MCT Blog (mct-master.github.io)
Unity Related Stuff:
#AltDevBlog » Unity3D coroutines in detail (jahej.com) - Good explanation on how Coroutines probably within Unity
Useful WebRTC links:
https://webrtc.org/
https://webrtc.googlesource.com/src/+/refs/heads/main/docs/native-code/index.md
https://webrtc.github.io/webrtc-org/native-code/development/
https://docs.microsoft.com/en-us/winrtc/getting-started
https://github.com/microsoft/winrtc
https://github.com/sipsorcery-org/sipsorcery/tree/710e3bb3bc9d787df45d97f1f80b4b700e4121f2
https://microsoft.github.io/MixedReality-WebRTC/manual/cs/cs.html
https://github.com/radioman/WebRtc.NET
https://www.c-sharpcorner.com/blogs/webrtc-web-real-time-communication
Really good description of how STUN/TURN works:
2008-08-cluecon-stun-turn-ice.pdf (viagenie.ca)
State of the Art Image Segmentation Survey
https://arxiv.org/pdf/2007.00047.pdf
NeRF rendering
https://www.matthewtancik.com/nerf
2210.13641.pdf (arxiv.org)
Camera Pose Determination
Cameras as Rays: Pose Estimation via Ray Diffusion (jasonyzhang.com)
DUSt3R: Geometric 3D Vision Made Easy (naverlabs.com)
Another great color resource:
Welcome to Bruce Lindbloom's Web Site
Article displaying differences between color distances (CIE1976 vs CIE2000)
https://making.lyst.com/2014/02/22/color-detection/
Good overview of color
Frequently asked questions about color (poynton.ca)
Survery of methods for computing surface normals from point clouds
Comparison of Surface Normal Estimation Methods for Range Sensing Applications
Good overview of PCA
A Step-by-Step Explanation of Principal Component Analysis (PCA) | Built In
ChaiScript: Easily embeddable scripting C-link language
http://chaiscript.com/index.html
Natural Docs (documentation system)
https://naturaldocs.org/
Useful features of C++17
https://www.bfilipek.com/2019/08/17smallercpp17features.html?m=1
https://www.linkedin.com/pulse/21-new-features-modern-c-use-your-project-vishal-chovatiya/
GitHub - hanickadot/compile-time-regular-expressions: Compile Time Regular Expression in C++
C++20 Draft
2019-02 Kona ISO C++ Committee Trip Report
C++ Algorithms
GitHub - HappyCerberus/book-cpp-algorithms: The Standard Algorithms in C++.
Good site for comparing possible software licenses:
https://tldrlegal.com/
Neat little camera
https://openmv.io/
Computing Barycentric coordinates of projected pointhttp://citeseerx.ist.psu.edu/viewdoc/download;jsessionid=ACDF35543780A74C1E36F71B687E2513?doi=10.1.1.165.1208&rep=rep1&type=pdf
Found this useful article on tensor decomposition (and tensors in general):
https://public.ca.sandia.gov/~tgkolda/pubs/bibtgkfiles/SAND2007-6702.pdf
They are super useful for handling any 3D dataset – point clouds, meshes, or voxels – without costly licenses.
Though you need knowledge and some coding skills to make these work (but perfect for lifelong learners).
𝟭. 𝗗𝗮𝘁𝗮 𝗖𝗮𝗽𝘁𝘂𝗿𝗲 𝗮𝗻𝗱 𝗔𝗰𝗾𝘂𝗶𝘀𝗶𝘁𝗶𝗼𝗻
• Meshroom: Reconstruct 3D scenes from images! Meshroom excels at photogrammetry, creating 3D models from a set of photographs. (https://alicevision.org/)
• PostShot: Create a 3D Gaussian Splatting Experience from Videos or Photos without having to handle the complexities of setting up CUDA or Python Environments. (https://www.jawset.com/)
𝟮. 𝗣𝗿𝗼𝗰𝗲𝘀𝘀𝗶𝗻𝗴 𝗮𝗻𝗱 𝗔𝗻𝗮𝗹𝘆𝘀𝗶𝘀 (𝗔𝗹𝗹 𝗙𝗼𝗿𝗺𝗮𝘁𝘀)
• Open3D: The Swiss Army knife! Open3D offers a comprehensive suite for data processing, visualization, and algorithms, ideal for beginners and experts alike. ( https://www.open3d.org/)
• PCL (Point Cloud Library): A large-scale project offering a robust set of algorithms for 3D data processing and analysis, including point clouds, meshes, and voxels. (PCL requires compilation, consider other options for quick use) ( https://pointclouds.org/)
𝟯. 𝗣𝗼𝗶𝗻𝘁 𝗖𝗹𝗼𝘂𝗱 𝗦𝗽𝗲𝗰𝗶𝗳𝗶𝗰
• CloudCompare: A versatile option for processing, editing, and visualizing point clouds. CloudCompare excels at many tasks. ( https://lnkd.in/eMUb2bXc)
• PDAL (Point Data Abstraction Library): A powerful library for efficiently translating and manipulating point cloud data in various formats. ( https://pdal.io/)
• lidR (R package): If you're an R user, lidR provides functionalities for manipulating and analyzing airborne LiDAR data for forestry applications. ( https://lnkd.in/e-rTe6BG)
4. Mesh Specific
• Trimesh: Work seamlessly with meshes! Trimesh offers tools for loading, analyzing, and visualizing 3D geometry, perfect for data pre-processing or 3D printing applications. ( https://trimesh.org/)
• MeshLab: MeshLab offers a powerful set of tools for editing and sculpting 3D meshes. It offers features for processing raw data produced by 3D digitization tools/devices and for preparing models for 3D printing (https://www.meshlab.net/)
Voxel Specific:
• PyTorch3D: While I do not necessarily recommend PyTorch3D, it facilitates the creation and training of 3D deep learning models applicable to voxel data. ( https://pytorch3d.org/)
• MagicaVoxel: Specifically designed to work with voxel data, MagicaVoxel offers functionalities for loading, manipulation, and analysis of voxel grids. (https://lnkd.in/euFeA29F)
General Visualization:
• ThreeJS + Potree: A beautiful set of WebGL visualization solutions. ThreeJS provides a comprehensive toolkit for creating Web-based 3D Experiences (https://threejs.org/) and Potree allows you to handle large 3D Point Clouds, with an Octree Structure optimized for Real-Time Experiences. (https://lnkd.in/eQ7u6bBD)
I hope this helps you!