Description

This course covers the basic concepts of computer vision. This course has two parts. The first part of classical computer vision discusses camera model, image processing, corner detection, edges, interest point, image warping, homography, model fitting, stereo and optical flow. The second part of modern computer vision with deep learning includes topics for convolutional neural network, transformer network, image classification, segmentation, and object detection.

Lectures

• Monday and Wednesday 1:30 pm to 2:45 pm, COB1 Room 120

Labs

• CSE-185-04L, Wednesday 7:30 pm to 10:20 pm, SE1 Room 100

Office hours and contact information

• Instructor: Meng Tang
• Email: mtang4@umcerced.edu
• Office hours: Online every Tuesday from 10AM to 11AM, Zoom link
• Teaching assistant: Weijie Lyu (wlyu3@ucmerced.edu)

Topics

• Camera model
• Image processing and filtering
• Corner detection
• Interest point
• Image warping
• Homograph and multi-view geometry
• Stereo and optical flow
• Convolutional neural network
• Segmentation
• Object detection
• Transformer network

Reading list

• Textbook
1. Computer Vision: Algorithms and Applications. Springer, New York, 2nd edition, 2022, Richard Szeliski (A free PDF version can be downloaded from https://szeliski.org/Book)
• Reference for background study: 
1. Multiple View Geometry in Computer Vision, Second Edition, Richard Hartley and Andrew Zisserman, Cambridge University Press, March 2004.

• All lecture slides will be available on the course website

Prerequisites

• Upper class standing. Basic knowledge of calculus, linear algebra

Grading

• 40% Assignments
• 20% Midterm exam
• 20% Final exam
• 20% Labs