timzhang642/3d-machine-learning

This repository is a curated reading list and resource guide for a research area called 3D Machine Learning, which sits at the intersection of computer vision, computer graphics, and the kind of AI that learns from data. The author started it as personal study notes and over time expanded it into a community reference. There is no software to install or run. The value is entirely in what it links to. The content is organized around different ways computers represent three-dimensional objects. A photo or video is flat, but a 3D object in a computer can be stored as a cloud of points, as a mesh of polygons (triangles that form a surface), as a solid block carved out of space, or reconstructed from multiple camera angles. Each of these representations has its own set of research papers, and the repository catalogs them by category. The sections cover quite a range: recognizing what a single object is, detecting multiple objects in a scene at once, figuring out where a body or object is oriented (pose estimation), rebuilding 3D structure from flat images, analyzing surface textures and materials, and understanding the layout of entire scenes. Each entry links to the original academic paper and often to the dataset used to test it. There is also a section listing university courses on the topic from schools like Stanford, MIT, and UCSD, along with links to public datasets researchers use for training and benchmarking. A Slack community is linked for people who want to discuss the field with others. This is most useful for someone who wants to get oriented in 3D AI research, figure out what papers exist on a specific problem, or find datasets to work with. It is not a beginner tutorial and does not explain the underlying math. The full README is longer than what was shown.