Computer Vision II

CSE 152B, Spring 2023

Instructor: Manmohan Chandraker
Email: mkchandraker [AT] ucsd [DOT] edu

Lectures: WF 6:30-7:50pm at WLH 2113
Instructor office hours: TBD at CSE 4122

TA: Vishal Vinod (vvinod@ucsd.edu)
TA office hours: TBD

Discussion section: TBD

Class discussion and message board: Piazza

Overview

Computer vision has made tremendous progress in recent years and is increasingly becoming a part of our daily lives. It seeks to analyze images to draw meaningful conclusions, which often requires drawing upon prior knowledge. This makes machine learning a useful tool, where the advent of deep neural networks has led to significant gains over the past five years. There is widespread acceptance that computer vision will play a major role in enabling technologies of the future, such as self-driving cars or augmented reality. The goal of the class is to equip students with the knowledge and skills to pursue higher studies or industry careers in modern computer vision.

Prerequisites

A background in linear algebra and calculus is required. Programming experience in Python is required. Courses that cover these might be Math 20F, CSE 100 or Math 176, CSE 101 or Math 188. Prior knowledge of basics in computer vision is recommended, as covered by CSE 152A. Students are encouraged to contact the instructor if unsure about meeting any criteria for enrollment.

Course Format and Requirements

The course will primarily involve lectures by the instructor. The goal of the course is to develop an understanding of the current state of computer vision and gain appreciation of its limits and potential. Students are encouraged to actively ask questions that the instructor may discuss.

Grades will be weighted as 35% for a final exam, 25% for a mid-term and 40% for assignments. There will be three homework assignments and a few ungraded quizzes.

Topics

The course will explore the fundamentals of diverse topics in computer vision and deep learning. We will do so by studying canonical vision tasks and machine learning tools:

Outline

Apr 05: Introduction Apr 12: Features and matching Apr 14: Learning features and distances Apr 19: Learning to match features Apr 21: Spatial transformers Apr 26: Optical flow Apr 26: Learning optical flow May 03: Structure from motion May 05: Learning structure from motion May 10: Mid-term review May 17: Face recognition: I May 19: Face recognition: II May 24: Human pose estimation May 26: Semantic segmentation: I May 31: Semantic segmentation: II Jun 02: Object detection: I Jun 07: Object detection: II Jun 09: Review

Resources


Manmohan Chandraker
Last modified: Tue, Apr 04, 2023