isabelle bloch, olivier colliot, and roberto m. cesar, jr. ieee trans. syst., man, cybern. b,...

34
On the Ternary Spatial Relation “Between” Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck February 22, 2011

Upload: eustace-robinson

Post on 17-Dec-2015

214 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

On the Ternary Spatial Relation “Between”Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr.IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006

Presented by Drew Buck

February 22, 2011

Page 2: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Our Goal

We are interested in answering two questions: What is the region of space located

between two objects A1 and A2?

To what degree is B between A1 and A2?

Page 3: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Outline

Existing Definitions Convex Hull Morphological Dilations Visibility Satisfaction Measure Properties Examples

Page 4: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Existing Definitions

Merriam-Webster Dictionary defines between as “In the time, space, or interval that separates.”

Some crisp approaches: A point is between two objects if it

belongs to a segment with endpoints in each of the objects.

A1 A2P

Page 5: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Existing Definitions

Another crisp approach: Using bounding spheres, B is between A1

and A2 if it intersects the line between the centroids of A1 and A2.

Page 6: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Existing Definitions

Fuzzy approaches: For all calculate the

angle θ at b between the segments [b, a1] and [b, a2]. Define a function μbetween(θ) to measure the degree to which b is between a1 and a2.

BbAaAa ,, 2211

[12] R. Krishnapuram, J. M. Keller, and Y. Ma, “Quantitative Analysis of Properties and Spatial Relations of Fuzzy Image Regions,” IEEE-Trans. Fuzzy Syst., vol. 1, no.3, pp. 222-233, Feb. 1993.

Page 7: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Existing Definitions

Using the histograms of forces: The spatial relationship between two

objects can be modeled with force histograms, which give a degree of support for the statement, “A is in direction θ from B.”

Page 8: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Existing Definitions

From P. Matsakis and S. Andréfouët, “The Fuzzy Line Between Among and Surround,” in Proc. IEEE FUZZ, 2002, pp. 1596-1601.

Page 9: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Existing Definitions

Limitations of this approach: Considers the union of objects, rather

than their individual areas.

Page 10: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Convex Hull

For any set X, its complement XC, and its convex hull CH(X), we define the region of space between objects A1 and A2 as β(A1,A2). )(\)()(, 2121212121 AAAACHAAAACHAA CC

CH

Page 11: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Convex Hull

Components of β(A1,A2) which are not adjacent to both A1 and A2 should be suppressed.

However, this leads to a continuity problem:

Page 12: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Convex Hull

Extension to the fuzzy case:

Recommended to chose a t-norm that satisfies the law of excluded middle, such as Lukasiewicz’ t-norm

)()()(,212121 AAAACH ccCHAA

).1,0max(),(],1,0[),( babatba

Page 13: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Non-connected Objects

If A1 can be decomposed into connected components we can define the region between A1 and A2 asand similarly if both objects are non-connected.

This is better than using CH(A1,A2 ) directly

i

iA1

i

i

i

i AAAAAA ),(),(, 212121

Page 14: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Morphological Dilations

Definition based on Dilation and Separation: Find a “seed” for β(A1,A2) by dilating both objects until

they meet and dilating the resulting intersection.where Dn denotes a dilation by a disk of radius n.

CCnnnDil AAADADDAA 212121 )()(),(

Page 15: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Wherefrom which connected components of the convex hull not adjacent to both sets are suppressed.

Morphological Dilations

Using watersheds and SKIZ (skeleton by influence zones): Reconstruction of β(A1,A2) by geodesic dilation.

)(

)(),(

)('

)('21

21

21

SKIZD

WSDAA

AACH

AACHsep

)(\)()(' 212121 AAAACHAACH

Page 16: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Morphological Dilations

Removing non-visible cavities We could use the convex hulls of A1 and

A2 independently, but this approach cannot handle imbricated objects.

Page 17: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Morphological Dilations

Fuzzy Directional Dilations: Compute the main direction α between A1 and A2 as the

average or maximum value of the histogram of angles.

Let Dα denote a dilation in direction α using a fuzzy structuring element, v.)](),([sup))(( yxvytxD

yv

),(,,),,()( 21221121),( 21 xAA uaaAaAaaah)'(max

)()(

),('

),(),(

21

21

21

AA

AAAA h

hH

Page 18: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Morphological Dilations

Given a fuzzy structuring element in the main direction from A1 to A2, we can define the area between A1 and A2 as

We can consider multiple values for the main direction by defining β as

CC AAADADAA 212121 )()(),(

)(),(),( 2121 AAAA

Page 19: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Morphological Dilations

Or use the histogram of angles directly as the fuzzy structuring element.

),()(),(

)(),(

1),(2

),(1

21

21

rvHrv

Hrv

AA

AA

CCvvFDil AAADADAA 2121211 )()(),(

12

Page 20: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Morphological Dilations

Alternate definition which removes concavities which are not facing each other

)()(

)()(),(

21

21212

22

11

ADAD

ADADAA

vv

vvFDil

Cvv

Cvv

CCvvFDil

ADAD

ADAD

AAADADAA

)()(

)()(

)()(),(

21

21

2121213

22

11

12

Page 21: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Visibility

Admissible segments: A segment [x1, x2] with x1 in A1 and x2 in A2

is admissible if it is contained in .

is defined as the union of all admissible segments.

CC AA 21

),( 21 AAAdm

Page 22: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Visibility

Fuzzy visibility: Consider a point P and the segments [a1, P] and

[P, a2] where a1 and a2 come from A1 and A2 respectively.

For each P, find the angle closest to π between any two admissible segments included in

CC AA 21

],[],,[,min)( 21min aPPaP

where [a1, P] and [P, a2] are semi-admissible segments

)())(,( min21 PfPAAFVisib

where f :[0, π] → [0,1] is a decreasing function

Page 23: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Visibility

Extension to fuzzy objects: Compute the fuzzy inclusion of each

segment and intersect with the best angle as before.

)(1),(1mininf]),([21

21 ],[21 yyaa AA

aayincl

],[],,[

]),([]),,([

),(),(,

max))(,(

21

21

2211

21

aPPa

aPPa

ASuppaASuppaft

PAAinclinclFVisib

Page 24: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Extended Objects

Objects with different spatial extensions: If A2 has infinite or near-infinite size with respect

to A1, approximate A2 by a segment u and dilate A1 in a direction orthogonal to u, limited to the closest half plane.

Page 25: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Extended Objects

Adding a visibility constraint Conjunctively combine projection with

admissible segments. Optionally consider only segments orthogonal to u.

Page 26: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Satisfaction Measure

Once β(A1,A2) has been determined, we want to know the overlap between β and B. Normalized Intersection:

Other possible measures:

B

BBS

),(1

0)(/)(sup1),(2 xxBS B

1),()(1mininf),(3 xxBS Bx

Page 27: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Satisfaction Measure

Degree of Inclusion:

Degree of Intersection:

Length of this interval gives information on the ambiguity of the relation.

)(1),(inf),( xxTBN Bx

)(),(sup),( xxtB Bx

Page 28: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Satisfaction Measure

If B is extended, we might want to know if B passes through β.

Let R will have two connected

components, R1 and R2. Degree to which B passes through β

should be high if B goes at least from a point in R1 to a point in R2.

)(\)( CoreSuppR

Page 29: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Satisfaction Measure

))((sup),)((supmin),(21

1 xBxBBM C

Rx

C

Rx

))()((sup),)()((supmin),(21

2 xCoreBxCoreBBM C

Rx

C

Rx

Page 30: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Properties

Page 31: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Examples

Anatomical descriptions of brain structures

Page 32: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Examples

Page 33: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Examples

Page 34: Isabelle Bloch, Olivier Colliot, and Roberto M. Cesar, Jr. IEEE Trans. Syst., Man, Cybern. B, Cybern., Vol. 36, No. 2, April 2006 Presented by Drew Buck

Conclusion

Possible to represent the complex spatial relationship, “between,” using mathematical morphology and visibility.

Many definitions exist, each with individual strengths and weaknesses.

Pick the representation most appropriate for the application.