01 Intro CS4495 8 18 03

Introduction: 

Introduction Jim Rehg CS 4495/7495 Computer Vision Lecture 1 August 18, 2003

Overview: 

Overview Introductory computer vision course in the Computational Perception and Robotics (CPR) curriculum Text: Computer Vision by Forsythe and Ponce Should be in the bookstore now Teaching Assistant: Mr. Jianxin Wu Course web site: Currently linked off my home page

CPR Curriculum Overview: 

CPR Curriculum Overview Pattern Recognition CS 4803 Computer Vision CS 4495/7495 Machine Learning CS 4640 Multi-view Geometry CS 8803 Computational Perception CS 7635 Spring ‘04 Spring ‘04 Intelligent Robotics CS 4630 Autonomous Robotics CS 4803/7630 Multi-Robot Systems CS 8803L Fall ‘04 Math for Comp. Perception CS 8803 Fall ‘03

Course Objectives: 

Course Objectives Broad introduction to research problems in computer vision, for example: How to recover depth information? How to recognize objects? Familiarity with standard vision results, such as: Multi-scale image representations Finding correspondences between images Experience programming vision algorithms using standard software environments VIPER toolkit and Intel IPL and OpenCV libraries Microsoft Visual C++ programming environment

Grading: 

Grading Two student populations: Students with a research focus Students who want a background in vision techniques (or an interesting technical elective) Two problem set tracks: Track 1: More in-depth and research-oriented Track 2: More broad and application-focused Graduate (7495) and undergraduate (4495) students will be graded separately.

Prerequisites: 

Prerequisites Linear algebra Least squares problems: Singular Value Decomposition: C++ Programming Microsoft Visual C++ A few tutorial sessions will be offered First is Viper/VC++, sometime next week Consider taking CS 8803 MCP this term if your math background is not strong.

Homework Assignments: 

Homework Assignments Most assignments will involve some programming and will be due in one week. Assignments will be available from course web site. I will try to stay 1-2 assignments ahead of what is currently due. There will be a take-home midterm and a final project.

Workload: 

Workload This course is going to be a lot of fun. It’s also going to be a lot of work. If you don’t have programming experience, then it’s REALLY going to be a lot of work. Think hard about your decision now, because the add/drop period is unfortunately very short. If you have questions or concerns, don’t hesitate to send me email or come by my office.

What is Computer Vision?: 

What is Computer Vision? The science of extracting information about the world from images. The most challenging mystery in Computer Science! Closely related fields: Computer Graphics Machine Learning Artificial Intelligence Image/Signal Processing

Characteristics of Human Vision: 

Characteristics of Human Vision Vision is a Constructive Process Your conscious perception of the visible world is an illusion manufactured by your brain (at great cost). Examples: brightness, color, and size constancy Vision Solves Specific Tasks in Specific Contexts Generality in visual skills tied directly to needs and context. Example: Thatcher illusion

Adelson Checkerboard Illusion: 

Adelson Checkerboard Illusion Perceived brightness is complex function of pixel values (Image courtesy of Ted Adelson) Brightness constancy problem

Color Constancy: 

Color Constancy Pixel color strongly affected by illumination. Perception of color constancy maintained by the brain Sunlight Fluorescent light (Images courtesy of David Heeger)

Size Constancy: 

Size Constancy Object size vs. object depth (Images copyright John H. Kranz, 1999)

Characteristics of Human Vision: 

Characteristics of Human Vision Vision is a Constructive Process Your conscious perception of the visible world is an illusion manufactured by your brain (at great cost). Examples: brightness, color, and size constancy Vision Solves Specific Tasks in Specific Contexts Generality in visual skills tied directly to needs and context. Example: Thatcher illusion

Thatcher Illusion: 

Thatcher Illusion (Due to P. Thompson)

Thatcher Illusion: 

Thatcher Illusion Face processing is sensitive to orientation.