Computer Vision January 2004 L1.1© 2004 by Davi Geiger

Introduction

Computer Vision January 2004 L1.2© 2004 by Davi Geiger

Vision

• ``to know what is where, by looking.’’ (Marr).

• Where• What

Computer Vision January 2004 L1.3© 2004 by Davi Geiger

Why is Vision Interesting?

• Psychology– ~ 35% of cerebral cortex is for vision.– Vision is how we experience the world.

• Engineering– Want machines to interact with world.– Digital images are everywhere.

Computer Vision January 2004 L1.4© 2004 by Davi Geiger

Vision is inferential: Light

(http://www-bcs.mit.edu/people/adelson/checkershadow_illusion.html)

Computer Vision January 2004 L1.5© 2004 by Davi Geiger

Vision is Inferential: Prior Knowledge

Computer Vision January 2004 L1.6© 2004 by Davi Geiger

Computer Vision January 2004 L1.7© 2004 by Davi Geiger

Computer Vision

• Inference Computation• Building machines that see• Modeling biological perception

Computer Vision January 2004 L1.8© 2004 by Davi Geiger

A Quick Tour of Computer Vision

Computer Vision January 2004 L1.9© 2004 by Davi Geiger

Boundary Detection

http://www.robots.ox.ac.uk/~vdg/dynamics.html

Computer Vision January 2004 L1.10© 2004 by Davi Geiger

Computer Vision January 2004 L1.11© 2004 by Davi Geiger

Boundary Detection

Finding the Corpus Callosum

(G. Hamarneh, T. McInerney, D. Terzopoulos)

Computer Vision January 2004 L1.12© 2004 by Davi Geiger

Tracking

Computer Vision January 2004 L1.13© 2004 by Davi Geiger

Tracking

Computer Vision January 2004 L1.14© 2004 by Davi Geiger

Tracking

Computer Vision January 2004 L1.15© 2004 by Davi Geiger

Tracking

Computer Vision January 2004 L1.16© 2004 by Davi Geiger

Tracking

Computer Vision January 2004 L1.17© 2004 by Davi Geiger

Stereo

Computer Vision January 2004 L1.18© 2004 by Davi Geiger

Stereo

http://www.magiceye.com/

Computer Vision January 2004 L1.19© 2004 by Davi Geiger

Motion

http://www.ai.mit.edu/courses/6.801/lect/lect01_darrell.pdf

Computer Vision January 2004 L1.20© 2004 by Davi Geiger

Motion - Application

(www.realviz.com)

Computer Vision January 2004 L1.21© 2004 by Davi Geiger

Pose Determination

Visually guided surgery

Computer Vision January 2004 L1.22© 2004 by Davi Geiger

Recognition - Shading

Lighting affects appearance

Computer Vision January 2004 L1.23© 2004 by Davi Geiger

Classification

(Funkhauser, Min, Kazhdan, Chen, Halderman, Dobkin, Jacobs)

Computer Vision January 2004 L1.24© 2004 by Davi Geiger

Vision depends on:

• Geometry• Physics• The nature of objects in the world (This is the hardest part).

Computer Vision January 2004 L1.25© 2004 by Davi Geiger

Approaches to Vision

Computer Vision January 2004 L1.26© 2004 by Davi Geiger

Modeling + Algorithms

• Build a simple model of the world (eg., flat, uniform intensity).• Find provably good algorithms.• Experiment on real world.• Update model.Problem: Too often models are simplistic or

intractable.

Computer Vision January 2004 L1.27© 2004 by Davi Geiger

Bayesian inference• Bayes law: P(A|B) = P(B|A)*P(A)/P(B).• P(world|image) = P(image|world)*P(world)/P(image)• P(image|world) is computer graphics

– Geometry of projection.– Physics of light and reflection.

• P(world) means modeling objects in world. Leads to statistical/learning approaches.Problem: Too often probabilities can’t be known and are invented.

Computer Vision January 2004 L1.28© 2004 by Davi Geiger

Engineering

• Focus on definite tasks with clear requirements.

• Try ideas based on theory and get experience about what works.

• Try to build reusable modules.Problem: Solutions that work under specific

conditions may not generalize.

Computer Vision January 2004 L1.30© 2004 by Davi Geiger

The State of Computer Vision

• Science– Study of intelligence seems to be hard.– Some interesting fundamental theory

about specific problems.– Limited insight into how these interact.

Computer Vision January 2004 L1.31© 2004 by Davi Geiger

The State of Computer Vision

• Technology– Interesting applications: inspection,

graphics, security, internet….– Some successful companies. Largest

~100-200 million in revenues. Many in-house applications.

– Future: growth in digital images exciting.

Computer Vision January 2004 L1.32© 2004 by Davi Geiger

Related Fields

• Graphics. “Vision is inverse graphics”.• Visual perception.• Neuroscience.• AI• Learning• Math: eg., geometry, stochastic processes.• Optimization.