| layout | default |
|---|
UC Berkeley, Spring 2026 Time: TuTh 15:30-16:59 Location: Hearst Mining 390 Ed: https://edstem.org/us/join/bX8tGp Gradescope Code: G66BDW
Instructors: Alyosha Efros, Angjoo Kanazawa
GSIs: Jiaxin Ge, Jorge Diaz Chao
Office hours - Location
- Alyosha Efros: After Lecture
- Angjoo Kanazawa: After Lecture
- Jiaxin Ge: Thursday 12-1pm BWW First Floor
- Jorge Diaz Chao: Monday 12-1pm BWW First Floor
Policy: To keep discussions organized, please do not email the instructor or GSIs directly. We are using Ed Discussion for communication (private or public) as much as possible.
Prerequisites: Students should have mastery over content from COMPSCI 189, COMPSCI 182, COMPSCI 180/280A. We will not be teaching basic image processing (convolution, gaussian smoothing, fourier transforms), basic deep learning (PyTorch, Jax, TensorFlow), stereo/homography.
- Lecture 1: (1/20) Course Intro. and Fundamentals of Image Formation
- Lecture 2: (1/22) Camera Calibration
- Lecture 3: (1/27) Calibration, Binocular Perspective
- Lecture 4: (1/29) Multi-View Geometry
- Lecture 5: (2/3) Biological Vision
- Lecture 6: (2/5) Perceptual Grouping and Object Recognition Part1
- Lecture 7: (2/10) Perceptual Grouping and Object Recognition Part2
- Lecture 8: (2/12) Modern Detection,Segmentation and Keypoints
- Lecture 9: (2/17) Attention
- Lecture 10: (2/19) ViT, Rope, Transformers
- Lecture 11: (2/24) Flow Matching
- Lecture 12: (2/26) Practicals of Flow Models
- Lecture 13: (3/3) Data (I)
- Lecture 14: (3/5) Data (II)
- Lecture 15: (3/10) Representation Learning
- Lecture 16: (3/12) Self-Supervised Learning
- Lecture 17: (3/17) Radiometry & Color
- Lecture 18: (3/31) Learning Correspondences
- Lecture 19: (4/2) Learning-based Multi-view 3D
- Lecture 20: (4/9) Single View 3D Prediction: Humans
- Lecture 21: (4/14) Single View 3D Prediction: Objects
- Lecture 22: (4/16) HDR and Image-Based Lighting
- Lecture 23: (4/21) 4D Vision
- Lecture 24: (4/23) World Models
- HW0: Due 01/27 at 11:59pm
- HW1: Due 02/10 at 11:59pm
- HW2: Due 02/24 at 11:59pm
- HW3: Due 03/17 at 11:59pm
- Final Project*