This preview shows pages 1–3. Sign up to view the full content.
1
CS252A, Fall 2010
Computer Vision I
Visual Tracking
Computer Vision I
CSE252A
Lecture 19
CS252A, Fall 2010
Computer Vision I
Announcements
• HW4 due tomorrow at midnight
• Final exam: Wed (12/8) 11:302:30 in WLH 2204
CS252A, Fall 2010
Computer Vision I
Main tracking notions
• State:
usually a finite number of parameters (a
vector) that characterizes the “state” (e.g.,
location, size, pose, deformation of thing being
tracked.
• Dynamics: How does the state change over time?
How is that changed constrained?
• Representation: How do you represent the thing
being tracked
• Prediction: Given the state at time
t1,
what is an
estimate of the state at time
t
?
• Correction: Given the predicted state at time
t
, and
a measurement at time
t
, update the state.
• Initialization – what is the state at time t=0?
CS252A, Fall 2010
Computer Vision I
What is state?
•
2D image location,
Φ
=(u,v)
•
Image location + scale
Φ
=(u,v,s)
•
Image location + scale + orientation
Φ
=(u,v,s,
θ
)
•
Affine transformation
•
3D pose
•
3D pose plus internal shape parameters (some may be
discrete).
– e.g., for a face, 3D pose +facial expression using FACS + eye
state (open/closed).
•
Collections of control points specifying a spline
•
Above, but for multiple objects (e.g. tracking a formation
of airplanes).
•
Augment above with temporal derivatives
CS252A, Fall 2010
Computer Vision I
Example: Blob Tracker
•
From input image
I(u,v)
(color?) at time
t
, create
a binary image by applying a function f(I(u,v)).
•
Clean up binary image using morphological
operators
•
Perform connected component exploration to
find “blobs.” – connected regions.
•
Compute their moments (mean and covariance
of coordinates of region), and use as state
•
Using state estimate from
t1
and
perform “data
association” to identify state in from t.
CS252A, Fall 2010
Computer Vision I
Tracking: Probabilistic framework
• Very general model:
– We assume there are moving objects, which have an
underlying state X
– There are measurements Y, some of which are
functions of this state
– There is a clock
• at each tick, the state changes
• at each tick, we get a new observation
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document2
CS252A, Fall 2010
Computer Vision I
Tracking State
•
Instead of “knowing state” at each instant, we treat the
state as random variables X
t
characterized by a pdf
P(X
t
)
or perhaps conditioned on other Random Variables e.g.,
P(X
t
 X
t1
),
etc.
•
The observation (measurement)
Y
t
is a random variable
conditioned on the state
P(Y
t
 X
t
)
•
Generally, we don’t observe the state – it’s hidden.
X
0
X
1
X
t1
X
t
X
t+1
Y
0
Y
1
Y
t1
Y
t
Y
t+1
…
CS252A, Fall 2010
Computer Vision I
Three main steps
We can try to express these conditional distributions parametrically, sample the
distribution, or estimate the mode.
CS252A, Fall 2010
This is the end of the preview. Sign up
to
access the rest of the document.
 Fall '08
 staff

Click to edit the document details