Computer Vision Spring 2012 15-385,-685 Instructor: S. Narasimhan Wean Hall 5409 T-R 10:30am – 11:50am.

Computer Vision

Spring 2012 15-385,-685

Instructor: S. Narasimhan

Wean Hall 5409

T-R 10:30am – 11:50am

A Picture is Worth 100 Words

A Picture is Worth 10,000 Words

A Picture is Worth a Million Words

A Picture is Worth a ...?

Necker’s Cube Reversal

A Picture is Worth a ...?

Checker Shadow Illusion – [E. H. Adelson]

A Picture is Worth a ...?

Checker Shadow Illusion – [E. H. Adelson]

Human Vision

• Can do amazing things like:

• Recognize people and objects• Navigate through obstacles• Understand mood in the scene• Imagine stories

• But still is not perfect:

• Suffers from Illusions• Ignores many details• Ambiguous description of the world• Doesn’t care about accuracy of world

Computer Vision

What we see

What a computer sees

Computer Vision

What we see

What a computer sees

What is Computer Vision?

• Inverse Optics

• Intelligent interpretation of Imagery

• Building a Visual Cortex

• No matter what your definition is…

– Vision is hard.

– But is fun...

Lighting

Scene

Camera

Computer

Scene Interpretation

Components of a Computer Vision System

Topics covered

Image Processing

Fourier TransformSampling, Convolution

Image enhancement Feature detection

Surface Reflectance

[CURET]

Lightness and Perception

Checker Shadow Illusion – [E. H. Adelson]

Understanding Optical Illusions

Which is bigger? Straight Lines?

Spinning Wheels?Dots White? Or Black?

3D from Shading

Shape from Shading Photometric Stereo

Cameras and their Optics

Today’s Digital Cameras

The Camera Obscura

Biological Cameras

Human Eye Mosquito Eye

Optical Flow

Tracking

Binocular Stereo

Range Scanning and Structured Light

Range Scanning and Structured Light

Microsoft Kinect

IR Camera

RGB Camera

IR LED Emitter

Statistical Techniques

Least Squares Fitting

Face detection

Face Recognition

• Principle Components Analysis (PCA)

• Face Recognition

Some Recent Trends in Vision

Novel Cameras and Displays

*** Topics change every year

• Graduate Level Computer Vision (Hebert, Fall)

• Computational Photography (Efros, Fall)

• Physics-based methods in Comp Vision (Narasimhan)

• Learning-based methods in Comp. Vision (Efros)

• Geometry-based methods in Comp. Vision (Hebert)

Advanced Related Courses at CMU

Course Logistics

• Class Notes (required)

• Text, Robot Vision, B.K.P.Horn, MIT Press (recommended)

• Supplementary Material (papers, tutorials)

Text and Reading

1/17/2012: Introduction and Course Fundamentals1/19/2012: Matlab Review

PART 1 : Signal and Image Processing1/24/2012 1D Signal Processing1/26/2012: 2D Image Processing [Project 1 OUT]1/31/2012: Image Pyramids and Sampling 2/2/2012: Edge Detection2/7/2012: Hough Transform

PART 2: Physics of the World2/9/2012: Surface appearance and BRDF2/14/2012: Photometric Stereo [Project 1 DUE, Project 2 OUT]2/16/2012: Shape from Shading2/21/2012: Direct and Global Illumination

PART 4 : 3D Geometry2/23/2012: Image Formation and Projection2/28/2012: Motion and Optical Flow3/1/2012: Lucas Kanade Tracking [Project 2 DUE Project 3 OUT]3/6/2012: Midterm Review3/8/2012: Midterm Exam

3/20/2012: Binocular Stereo 13/22/2012: Binocular Stereo 2 [Project 3 DUE, Project 4 OUT]3/27/2012: Structured Light and Range Imaging

Course Schedule

PART 4 : Statistical Techniques3/29/2012: Feature Detection 14/03/2012: Classification 14/05/2012: Classification 24/10/2012: Principle Components Analysis [Project 4 DUE]4/12/2012: Applications of PCA [Project 5 OUT]

[Grad project description due]

PART 6: Trends and Challenges in Vision Research4/17/2012: Image Based Rendering4/24/2012: Novel Cameras and Displays4/26/2012: Optical Illusions5/1/2012: Open challenges in vision research [Project 5 DUE]

5/3/2012: Project presentations by undergraduate students5/8/2012: Project presentations by graduate students [Grad Project 6 DUE]5/13/2012: Final Grades Due

Course Schedule

*** Use as a guide…changes possible

• Basic Linear Algebra, Probability, Calculus Required

• Basic Data structures/Programming knowledge

• No Prior knowledge of Computer Vision Required

Prerequisites

• FIVE Projects – 90 % (15%, 15%, 20%, 20%, 20%)

• ONE Midterm – 10 %

• ONE Extra Project for Graduate Students – 20 %

• Most projects include analytic and programming parts.

• All projects must be done individually.

• Programming Environment – Matlab.

• Projects due before midnight on due-date.

• Written parts due in class on the due-date.

• 3 Late Days for the semester. No more extensions.

• Class attendance – 5 % extra credit

Grading

Office Hours

Narasimhan: Smith Hall 223, By Appointment Email: srinivas@cs.cmu.edu

Supreeth Achar: Wednesdays 6:00pm – 8:00pm Email: supreeth@cmu.edu

Gunhee Kim: Thursdays, Thursdays 6:00pm – 8:00pm Email: gunhee@cs.cmu.edu

• Technical Questions: Post on bboard. TAs will answer.