Introduction To Data Association
Introduction To Data Association
CSE598C
track 1
?
track 2 observations
track 1
?
track 2 observations
d1
d3
?
track 1 d2
d5
d4
track 2 observations
?
track 1 d2
track 2 observations
?
track 1
gating ?
region 1
track 2 observations
r r constant position
prediction
pk-1 pk prediction
pk + (pk-pk-1)
(pk-pk-1)
typically, gating
region can be smaller
Robert Collins
CSE598C
Global Nearest Neighbor (GNN)
Evaluate each observation in track gating region.
Choose “best” one to incorporate into track.
ai1
o2 1 3.0
o1
2 5.0
track1 o3 3 6.0
o4 4 9.0
A
2 * area(A and B)
score =
area(A) + area(B)
B
constant velocity
assumes V(t) = V(t+1)
V(t) V(t+1)
2 * area(A and B)
score =
area(A) + area(B)
B
Update appearance
template of blobs
A B C D
(A,C) = 0.39
A -> D
(A,D) = 2.03
1 2 3 4 5
1 0.95 0.76 0.62 0.41 0.06
2 0.23 0.46 0.79 0.94 0.35
3 0.61 0.02 0.92 0.92 0.81
4 0.49 0.82 0.74 0.41 0.01
5 0.89 0.44 0.18 0.89 0.14
subject to:
subject to:
subject to:
source sink
weighted Possible solution methods:
links Ford-Fulkerson algorithm?
Robert Collins
CSE598C
Assignment Problem
However, it is not a general max-flow problem, due to the constraints
N N
maximize: E=
subject to:
step 3: Draw as few row, col lines as possible to cover all the zeros. Note, in the
graph that you are implicitly working with, this is a minimum vertex cover of the
subgraph formed by zero-weight edges.
Robert Collins
CSE598C
Hungarian Algorithm
step 4: From the elements that are left, find the lowest value. Subtract this from all
elements that are not struck. Add this to elements that are present at the intersection
of two lines. Leave other elements unchanged.
Robert Collins
CSE598C
Hungarian Algorithm
score: 4.26
permutation matrix computed
by Hungarian Algorithm
k1
k2
ac
ac
tr
tr
1 3.0 0
2 5.0 0
pad with array of small
3 6.0 1.0 5x3 random numbers to get
4 9.0 8.0 a square score matrix.
5 0 3.0
k1
k2
ac
ac
tr
tr
1 0 0
Square-matrix 2 0 0 5x3
3 1 0
assignment 4 0 1 ignore whatever
5 0 0 happens in here
Robert Collins
CSE598C
K-Best Assignment
score: 4.26
permutation matrix
computed by SoftAssign
constraints
~(1,1) (1,1)(4,2),~(2,3)
(1,1),~(4,2) (1,1)(4,2),(2,3),~(5,4)
0.95 0.76 0.62 0.41 0.06 0.95 0.76 0.62 0.41 0.06
0.23 0.46 0.79 0.94 0.35 0.23 0.46 0.79 0.94 0.35
0.61 0.02 0.92 0.92 0.81 0.95 0.76 0.62 0.41 0.06 0.61 0.02 0.92 0.92 0.81 0.95 0.76 0.62 0.41 0.06
0.49 0.82 0.74 0.41 0.01 0.23 0.46 0.79 0.94 0.35 0.49 0.82 0.74 0.41 0.01 0.23 0.46 0.79 0.94 0.35
0.89 0.44 0.18 0.89 0.14 0.61 0.02 0.92 0.92 0.81 0.61 0.02 0.92 0.92 0.81
0.89 0.44 0.18 0.89 0.14
0.49 0.82 0.74 0.41 0.01 0.49 0.82 0.74 0.41 0.01
0.89 0.44 0.18 0.89 0.14 0.89 0.44 0.18 0.89 0.14
(5,1)(1,2),(4,3),(2,4),(3,5) (1,1)(4,2),(3,3),(5,4),(2,5)
score 4.14 (1,1)(5,2),(4,3),(2,4),(3,5) score 3.93 (1,1)(4,2),(2,3),(3,4),(5,5)
score 3.88 score 3.62
constraints
~(1,1),~(5,1) ~(1,1)(5,1),(1,2),~(4,3)
~(1,1),(5,1),~(1,2) ~(1,1)(5,1),(1,2),(4,3),~(2,4)
0.95 0.76 0.62 0.41 0.06 0.95 0.76 0.62 0.41 0.06
0.23 0.46 0.79 0.94 0.35 0.23 0.46 0.79 0.94 0.35
0.61 0.02 0.92 0.92 0.81 0.95 0.76 0.62 0.41 0.06 0.61 0.02 0.92 0.92 0.81 0.95 0.76 0.62 0.41 0.06
0.49 0.82 0.74 0.41 0.01 0.23 0.46 0.79 0.94 0.35 0.49 0.82 0.74 0.41 0.01 0.23 0.46 0.79 0.94 0.35
0.89 0.44 0.18 0.89 0.14 0.61 0.02 0.92 0.92 0.81 0.61 0.02 0.92 0.92 0.81
0.89 0.44 0.18 0.89 0.14
0.49 0.82 0.74 0.41 0.01 0.49 0.82 0.74 0.41 0.01
0.89 0.44 0.18 0.89 0.14 0.89 0.44 0.18 0.89 0.14
(4,1)(1,2),(2,3),(5,4),(3,5) (5,1)(1,2),(2,3),(4,4),(3,5)
score 3.74 (5,1)(4,2),(1,3),(2,4),(3,5) score 3.66 (5,1)(1,2),(4,3),(3,4),(2,5)
score 4.08 score 3.66
constraints
~(3,5) (3,5)(5,4),~(2,3)
(3,5),~(5,4) (3,5)(5,4),(2,3),~(4,2)
0.95 0.76 0.62 0.41 0.06 0.95 0.76 0.62 0.41 0.06
0.23 0.46 0.79 0.94 0.35 0.23 0.46 0.79 0.94 0.35
0.61 0.02 0.92 0.92 0.81 0.95 0.76 0.62 0.41 0.06 0.61 0.02 0.92 0.92 0.81 0.95 0.76 0.62 0.41 0.06
0.49 0.82 0.74 0.41 0.01 0.23 0.46 0.79 0.94 0.35 0.49 0.82 0.74 0.41 0.01 0.23 0.46 0.79 0.94 0.35
0.89 0.44 0.18 0.89 0.14 0.61 0.02 0.92 0.92 0.81 0.61 0.02 0.92 0.92 0.81
0.89 0.44 0.18 0.89 0.14
0.49 0.82 0.74 0.41 0.01 0.49 0.82 0.74 0.41 0.01
0.89 0.44 0.18 0.89 0.14 0.89 0.44 0.18 0.89 0.14
(1,1)(4,2),(3,3),(5,4),(2,5) (1,1)(2,2),(4,3),(5,4),(3,5)
score 3.93 (5,1)(1,2),(4,3),(2,4),(3,5) score 3.85 (4,1)(1,2),(2,3),(5,4),(3,5)
score 4.14 score 3.74