csc508 computer vision introduction. csc508 what is it? a ripe lemon

45
CSC508 Computer Vision Introduction

Upload: gerald-phelps

Post on 17-Jan-2016

220 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Computer Vision

Introduction

Page 2: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

A ripe lemon

Page 3: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

An unripe lemon

Page 4: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

A ripe lemon

Page 5: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

A ripe lemon

Page 6: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

A ripe lemon

Page 7: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

A ripe orange

Page 8: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

An unripe orange

Page 9: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

A ripe orange?

Page 10: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

A tennis ball

Page 11: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

A baseball

Page 12: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

A “frisbee”

Page 13: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

A ripe lemon

Page 14: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

An unripe lemon

Page 15: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

A ripe orange?(see next slide)

Page 16: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What are they?

A ripe tangerine and a ripe orange

Page 17: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What is it?

A flounder

Page 18: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Meet Lenna

Page 19: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Still Lenna?

Page 20: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Still Lenna?

Page 21: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Still Lenna?

Page 22: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Still Lenna?

Page 23: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Still Lenna?

Page 24: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Illusions

• And now for something completely different…

Kanizsa Triangle (and variants)

Page 25: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What do we process?

• Shape• Color• Texture • Shading• Gradients (edges)• Context• Transformations

– Rotation– Translation– Stretching

Page 26: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

How do we process visual scenes?

• Preprocessing– Image conditioning

• Low Level Vision– Feature extraction

• Mid Level Vision– Feature description

• High Level Vision– Object recognition

Page 27: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

What we (the human viewer) “see”

Page 28: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Mechanization of Processing

• Cones: process color (3 types)• Primarily in the foveal region of the retina

• Rods: process shades of gray• Primarily in the peripheral region of the retina

Page 29: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

How should a computer process visual scenes?

• As it turns out, mimicking a biological system in circuits and software is extremely difficult– Cameras are not as sophisticated as the eye– Processors/software are not as sophisticated as the

brain

Page 30: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

156 145 137 146 163 151

150 173 171 177 174 171

104 132 133 142 164 165

86 116 164 162 155 152

95 121 150 137 123 136

123 128 141 129 126 129

Here’s what the computer “sees”

177 167 153 141 145 126

144 153 160 172 173 158

116 138 138 140 152 151

65 96 131 140 148 145

71 98 147 127 120 131

99 105 144 116 117 123

238 237 234 227 223 216

229 227 224 220 225 221

205 212 221 220 225 220

177 192 213 207 212 217

164 180 211 208 209 215

190 194 220 212 210 219

Page 31: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Here’s what the computer “sees”

Page 32: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Biological System is Restrictive

• We can improve a computer vision system by including multi-spectral data

Page 33: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Visible Light (gray scale)

Page 34: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Visible Light (color)

Page 35: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Infrared Image

“Night Vision”

Page 36: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Radar Image

“Through Smoke and Clouds”

Page 37: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Magnetic Resonance Imaging (MRI)

Page 38: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

X-Ray

Page 39: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Ultrasound

Page 40: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

How should a computer process visual scenes?

• Preprocessing– Image conditioning

• Low Level Vision– Feature extraction

• Mid Level Vision– Feature description

• High Level Vision– Object recognition

• Multi-spectral fusion– Combine images from different cameras

Page 41: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Plans for the Class

• We will be studying selected algorithms from each of the areas– We will be…

• …writing code to implement the algorithms

• …experimenting with parameters to understand the sensitivity of the algorithms

– There’s no better way to learn this topic than by doing

Page 42: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Programming

• Choose any language you like– Some languages provide image I/O, some do not– For those that do not, you’ll have to figure out how to

read/write images

• One option: ImageJ– Free from http://rsb.info.nih.gov/ij/– Full featured image processing tool kit– Extendable through plugins written in Java– Note that if you choose this option, you may NOT use the

provided processing functions. ImageJ will only be used to perform image I/O and display

Page 43: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Reference Books• Don’t feel obligated to run out and buy these for the class!

– Feature Extraction & Image Processing• Nixon and Aguado – Newnes Press

– Vision Science: Photons to Phenomenology• Stephen E. Palmer – MIT Press

– Vision• David Marr – Freeman Press

– Eye, Brain and Vision• David H. Hubel – Scientific American Library

– Computer Vision• Linda G. Shapiro, George C. Stockman – Prentice Hall

– Robot Vision• Berthhold Klaus Paul Horn – MIT Press

– Digital Image Processing• Rafael C. Gonzalez, Richard E. Woods – Prentice Hall

– Visual Intelligence• Donald D. Hoffman – W. W. Norton & Company

Page 44: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Reading

• Chapter 1

• Chapter 2

• Chapter 7

Page 45: CSC508 Computer Vision Introduction. CSC508 What is it? A ripe lemon

CSC508

Homework

• Describe a task that you think might benefit from computer vision (i.e. one that does not currently use visual processing)– How is the task currently performed?– How could computer vision make it better?– Do you think it’s feasible to implement?– For example: a few years back someone felt that a

computer mouse might be well suited to visual processing

• Due next class meeting