Theory of Fusion 1 Theory of Fusion sensor

  • Slides: 24
Download presentation
Theory of Fusion 1

Theory of Fusion 1

Theory of Fusion sensor fusion, data fusion, classifier fusion, ……. feature fusion, decision fusion,

Theory of Fusion sensor fusion, data fusion, classifier fusion, ……. feature fusion, decision fusion, agent fusion, …. . blank fusion…. . What IS fusion ? ? 2

Theory of Fusion sensor fusion, data fusion, classifier fusion, ……. feature fusion, decision fusion,

Theory of Fusion sensor fusion, data fusion, classifier fusion, ……. feature fusion, decision fusion, agent fusion, …. . blank fusion…. . What IS fusion ? ? Fusion is the process of combining of several objects (from the same category) to a single object (in the same category) for better performance. 3

Theory of Fusion sensor fusion, data fusion, classifier fusion, ……. feature fusion, decision fusion,

Theory of Fusion sensor fusion, data fusion, classifier fusion, ……. feature fusion, decision fusion, agent fusion, …. . blank fusion…. . What IS fusion ? ? Fusion is the process of combining of several objects (from the same category) to a single object (in the same category) for better performance. Objects – depends on the problem Category – an area of mathematical research started in 1940’s Combining – a rule or rules created by designer Performance – specified by designer Better – than the performance of any single object 4

Theory of Fusion 5

Theory of Fusion 5

Theory of Fusion - objects 6

Theory of Fusion - objects 6

Theory of Fusion – several objects 7

Theory of Fusion – several objects 7

Theory of Fusion - combine 8

Theory of Fusion - combine 8

Theory of Fusion – performance 9

Theory of Fusion – performance 9

Theory of Fusion – better performance 10

Theory of Fusion – better performance 10

Theory of Fusion - Context Event sets Systems A B Performance is the Affordance

Theory of Fusion - Context Event sets Systems A B Performance is the Affordance (? ? ) Concept encoding sets Is event context for event ? System A will say is context for w. r. t. the performance if there exists a concept combining rule such that (assume larger is better). That is , System B will say is context for w. r. t. the performance if there exists a concept combining rule such that 11

SITUATION EXAMPLE Entities are people 12

SITUATION EXAMPLE Entities are people 12

Objects are singleton sets of people Queries q 1: Is X a male? 13

Objects are singleton sets of people Queries q 1: Is X a male? 13

Queries q 1: Is X a male? 14

Queries q 1: Is X a male? 14

Queries q 1: Is X a male? q 1: Is X not a male?

Queries q 1: Is X a male? q 1: Is X not a male? 15

Queries q 1: Is X a male? q 2: Is X a sibling of

Queries q 1: Is X a male? q 2: Is X a sibling of Y? 16

A B C D A no no B no yes yes C no yes

A B C D A no no B no yes yes C no yes yes D no yes Yes yes 17

Queries q 1: Is X a male? q 2: Is X a sibling of

Queries q 1: Is X a male? q 2: Is X a sibling of Y? 18

Queries q 1: Is X a male? q 2: Is X a sibling of

Queries q 1: Is X a male? q 2: Is X a sibling of Y? 19

Queries q 1: Is X a male? q 2: Is X a sibling of

Queries q 1: Is X a male? q 2: Is X a sibling of Y? q 3: Is X a parent of Y? 20

Queries q 1: Is X a male? q 2: Is X a sibling of

Queries q 1: Is X a male? q 2: Is X a sibling of Y? q 3: Is X a parent of Y? What is “composition” of with ? 21

Queries q 1: Is X a male? q 2: Is X a sibling of

Queries q 1: Is X a male? q 2: Is X a sibling of Y? q 3: Is X a parent of Y? What is “composition” of with ? 22

23

23

24

24