课程定位
CS 348N 关注如何用机器学习生成和操作高质量 3D 模型与场景。课程覆盖几何表示、3D 数据集、点云/网格网络、neural implicit、3D 生成模型、条件生成和评估指标,是 3D 生成/图形方向的核心课。
主要内容
- 几何表示:点云、网格、体素、SDF、NeRF
- 3D 数据:ShapeNet、Objaverse、场景数据
- 点云/网格网络:PointNet、GCN、MeshCNN
- Neural implicit:SDF、occupancy、NeRF
- 3D 生成:VAE/GAN/diffusion/autoregressive
- 条件生成:文本/图像到 3D
- 评估与合成数据
先修
CS 148;深度学习基础;CS 229 推荐。
考核与工作量
作业 + 项目。具体以当季官网为准。
适合谁
3D 生成、图形学、AR/VR、机器人仿真方向。
学习建议
先学 NeRF 和 diffusion,再回头看 GAN/VAE 的 3D 版本。配合 CS 468。
相关课程
官方课程描述(英文原文)
Course Description: Generation of high-quality 3D models and scenes by leveraging machine learning tools and approaches. Survey of geometry representations. Public 3D object and scene data sets. Neural architectures for geometry, including deep architectures for point clouds and meshes. Generative models for 3D: autoencoders, GANs, neural implicits, neural ODEs, autoregressive models. Conditional generation based on images or partial geometry. Variation generation. Evaluation metrics for content generation. Use of synthetic data in ML training pipelines. Prerequisites: CS148 and the rudiments of deep learning. Recommended: CS229.