Useful Links:

• Visual SLAM Roadmap
• First Principles of Computer Vision
• CMU Course: 16-825: Learning for 3D Vision
• An Invitation to 3D Vision: A Tutorial for Everyone
• 3D Vision (UIUC), 3D Vision (CS 598) – Fall 2021
• Multiview 3D Geometry in Computer Vision (UMN), Spring 2018 CSCI 5980 Multiview 3D Geometry in Computer Vision

Topics in 3D Vision:

• 3D Reconstruction: SfM/SLAM, Multi-view Stereo, RGB-D Fusion.
• 3D Scene Understanding: plane, normal, depth
• Depth Estimation, Room Layout Estimation, NeRFs, Inverse Rendering,
• Computer Graphics and Computational Photography

Also includes:

• Representation Learning
• Image and Video Synthesis
• Vision for Robotics and Autonomous Vehicles

Introduction

3D Scene Understanding

• Task: Holistic reasoning of everything that is in the scene
• Involves many semantic tasks: Semantic Segmentation, Instance-level segmentation, Segmentation of the scene into semantic labels.
• Tracking

For example, from a short (≈10s) video sequence to infer:

• Geometric properties, e.g., street orientation
• Topological properties, e.g., number of intersecting streets
• Semantic activities, e.g., traffic situations at an intersection
• 3D objects, e.g., cars