CS194-26/294-26: Intro to Computer Vision and Computational Photography Computer Science Division University of California Berkeley

INSTRUCTOR: Alexei (Alyosha) Efros (Office hours: after lecture)
GSI: Zhe Cao (Office hours: 9 - 10 AM Fri)
Tutors: Chenyue Cai (Office hours: 8 - 9 AM Mon, Email: chenyue_cai[at]berkeley[at]edu), Kenny Chen (Office hours: 9 - 10 M Wed, Email: kenchen10[at]berkeley[at]edu), Yibin Li (Office hours: 11 - 12 PM Thurs, Email: liyibin516[at]berkeley[at]edu), and Violet Yao (Office hours: 1 - 2 PM Tuesday, Email: violetyao[at]berkeley[at]edu).
Reader: Justin Wang (Email: justinyiwang[at]berkeley[at]edu)
UNIVERSITY UNITS: 4
SEMESTER: Fall 2020
WEB PAGE: http://inst.eecs.berkeley.edu/~cs194-26/fa20/
Q&A: Piazza Course Website
LOCATION: Remote (Zoom links are posted on Piazza)
TIME : MW 5:00 PM-6:30 PM

PREREQUISITES:
This is a heavily project-oriented class, therefore good programming proficiency (at least CS61B) is absolutely essential. Moreover, familiarity with linear algebra (MATH 54 or EE16A/B or Gilbert Strang's online class) and calculus are vital. Experience with neural networks (e.g. CS189) is a plus. For these taking CS294-26, consent of instructor is required to register (please sign up on the waitlist first).

COURSE DESCRIPTION:
The aim of this advanced undergraduate course is to introduce students to computing with visual data (images and video). We will cover acquisition, representation, and manipulation of visual information from digital photographs (image processing), image analysis and visual understanding (computer vision), and image synthesis (computational photography). Key algorithms will be presented, ranging from classical (e.g. Gaussian and Laplacian Pyramids) to contemporary (e.g. ConvNets, GANs), with an emphasis on using these techniques to build practical systems. This hands-on emphasis will be reflected in the programming assignments, in which students will have the opportunity to acquire their own images and develop, largely from scratch, the image analysis and synthesis tools for solving applications.

PROGRAMMING ASSIGNMENTS:

Project 1: Images of the Russian Empire -- colorizing the Prokudin-Gorskii photo collection

See student submissions here

Class Choice Awards: Saurav Mittal

Project 2: Fun with Filters and Frequencies

See student submissions here

Class Choice Awards: Sara Wang

Project 3: Face Morphing and Modelling a Photo Collection

See student submissions here

Class Choice Awards: Saurav Mittal, Apollo Thomopoulos

Project 4: Face Keypoint Detection with Neural Networks

See student submissions here

Kaggle Winner: Kevin Lin
Class Choice Awards: Saurav Mittal

Project 5: (Auto)stitching and photo mosaics

See student submissions partA partB

Class Choice Awards: Apollo Thomopoulos

Final Project

See student submissions pre-canned own proposed

TEXTBOOK:
We will be loosely using the new 2nd edition of Rick Szeliski's Computer Vision textbook. The latest draft is available off the textbook's website. If you find a bug or a typo, please e-mail Rick for a chance to get an acknowledgement in the finished book! The first edition is still available at the bookstore, but it's missing some important things, like discussion of Convolutional Neural Networks.

There is a number of other fine texts that you can use for general reference:

Computer Vision: A Modern Approach (2^nd edition), Forsyth and Ponce (classic computer vision text)
Vision Science: Photons to Phenomenology, Stephen Palmer (great book on human visual perception)
Digital Image Processing, 2nd edition, Gonzalez and Woods (a good general image processing text)
Linear Algebra and its Applications, Gilbert Strang (a truly wonderful book on linear algebra)

CLASS NOTES
The instructor is extremely grateful to a large number of researchers for making their slides available for use in this course. Steve Seitz and Rick Szeliski have been particularly kind in letting me use their wonderful lecture notes. In addition, I would like to thank Paul Debevec, Stephen Palmer, Paul Heckbert, David Forsyth, Steve Marschner and others, as noted in the slides. The instructor gladly gives permission to use and modify any of the slides for academic and research purposes. However, please do also acknowledge the original sources where appropriate.

TENTATIVE CLASS SCHEDULE:

CLASS DATE	TOPICS	Material
Aug 26	Introduction	Slides: pdf, pptx How Photography Became an Art Form by Aaron Hertzmann
Aug 31	Capturing Light... in man and machine	Slides: pdf, pptx
Sep 2	Point Processing	Szeliski Ch. 2 Slides: pdf, ppt
Sep 9	Image Processing Filtering I	Start Szeliski Ch 3 Slides: pdf, ppt Reinhard et al., Color Transfer Between Images, IEEE Computer Graphics and Applications, 2001
Sep 14	Image Processing Filtering II	Continue Szeliski Ch 3 Slides: pdf, ppt
Sep 17	The Frequency Domain	Szeliski Ch 3.4 Slides: pdf, ppt
Sep 21	Pyramid Blending, Templates, NL Filters	Szeliski 3.5, 3.3 Slides: pdf, ppt Burt and Adelson, A multiresolution spline with application to image mosaics, ACM ToG, 1983
Sep 23	Spatial Frequencies and Human Perception	Slides: pdf, ppt
Sep 28	Image Morphing I	Szeliski 3.6 Slides: pdf, ppt
Sep 30	Image Morphing II	Continue Szeliski Ch 3
Oct 5	Data-driven Methods: Faces	Slides: pdf, ppt Galton, "Composite portraits made by combining those of many different persons into a single figure.", Nature, 1878 Rowland and Ferrett, "Manipulating Facial Appearance through Shape and Color", CG&A, 1995 Blanz and Vetter, "A Morphable Model for the Synthesis of 3D Faces", SIGGRAPH 1999 Cootes, Edwards, and Taylor, "Active Appearance Models", ECCV 1998
Oct 7	Data-driven Methods: Video Textures	Slides: pdf, ppt Schodl et al., Video Textures SIGGRAPH' 00 Efros and Leung, Texture Synthesis by Non-parametric Sampling ICCV'99 Efros and Freeman, Image Quilting for Texture Synthesis and Transfer, SIGGRAPH'01 Hertzmann et al. Image Analogies, SIGGRAPH 2001.
Oct 12	Visual Texture (in human and machine)	Slides: pdf, ppt
Oct 19	Feature Learning with Neural Networks	Slides: pdf, ppt
Oct 22	Convolutional Neural Networks I	Slides: pdf, ppt
Oct 26	Convolutional Neural Networks II	Slides: pdf, ppt Colorful Image Colorization Zhang et al., 2016 Learning Visual Similarity for Product Design with Convolutional Neural Networks, Bell et al, 2015
Oct 28	ConvNets for Image Synthesis	Slides: pdf, ppt Szeliski 5.5.3 Pix2Pix Isola et al., 2017 Everybody Dance Now Chan et al., 2018 CycleGAN Zhu et al., 2017 Neural Style Transfer, Gatys et al, 2015
Nov 03	The Camera	Slides: pdf, ppt
Nov 05	Modeling Light	Slides: pdf, ppt Szeliski 14.3
Nov 09	Homographies and Mosaics	Slides: pdf, ppt Szeliski Ch 8
Nov 16	Automatic Image Alignment	Slides: pdf, ppt
Nov 23	Scene Modeling for a Single View	Slides: pdf, ppt
Nov 30	Multiview Geometry: Stereo & Structure from Motion	Slides: pdf, ppt
Dec 2	What Makes a Great Picture?	Slides: pdf, ppt

CAMERAS:
Although it is not required, students are highly encouraged to obtain a digital camera for use in the course.

METHOD OF EVALUATION:
Grading will be based on a set of programming and written assignments (60%), 6-7 (with 1 drop) in-class pop quizzes (20%) and a final project (20%). For the programming assignments, students will be allowed a total of 5 (five) late days per semester; each additional late day will incur a 10% penalty.

Students taking CS294-26 will also be required to submit a conference-style paper describing their final project.

PROGRAMMING RESOURCES:
Students will be encouraged to use either MATLAB (with the Image Processing Toolkit) or Python (with either scikit-image or opencv) as their primary computing platform. Specific libraries in both languages offer tons of build-in image processing functions. Here is a link to some useful MATLAB and Python resources compiled for this class.

PREVIOUS OFFERINGS OF THIS COURSE:
Previous offerings of this course can be found here.

SIMILAR COURSES IN OTHER UNIVERSITIES:

• Computational Photography (Hoiem, UIUC)
• Computational Photography (Hays, Brown)
• Digital and Computational Photography (Durand, MIT)
• Computer Vision (Seitz & Szeliski, UWashington)

Page design courtesy of Doug James