Computer Vision

Question 1

Computer vision is most easily done as an entirely bottom up (starting with raw signal and inferring from there) task.

Answer

true

Question 2

Bayesian inference provides a means for integrating prior information with empirical information gathered from image data.

Answer

true

Question 3

Which is a big problem that must be tackled in computer vision?

Accepted Answer

Decision making under uncertainty

Answer

Decision making under certainty

Question 4

How does bayesian inference intepret probability?

Accepted Answer

Degree of belief

Answer

Probability of occuring

Question 5

If H is hypothesis and D is data what form would a human experts knowledge be in?

Accepted Answer

P(D|H)

Answer

P(H|D)

Question 6

In bayesian inference, to update our decision we use;

Accepted Answer

Latest posterior as new prior

Answer

Latest prior as new posterior

Question 7

What is one difficulty with computer vision tasks when using prior probabilities in bayesian inference

Accepted Answer

There are often no useful/strong priors

Answer

There are too many priors

Question 8

For reliable recognition we want:

Accepted Answer

Between-class variability to be larger than within-class variability

Answer

Within-class variability to be larger than between-class variability

Question 9

The posterior probability (likelihood of being a class given the data) is:

Accepted Answer

P(Ck|x)

Answer

P(x|Ck)

Create a free account to finish this quiz