Cellular feature
?Download as PPTX, PDF?
0 likes?436 views
This document describes a cellular automaton structure called cellular feature for extracting linear and curved features from images. Cellular feature represents the orientation and location of edges in an image through a local iterative computation method. Neighboring cells are defined and the integration algorithm performs parallel, local computations to integrate edge features on a digital plane.
Cellular feature
- 2. 2003.3.28/R.Oka 2 A single global feature of image = A whole image The global feature should be robust for geometrical variations Global feature extraction
- 3. 2003.3.28/R.Oka 3 Form and edge feature form integrated edge feature at the center of figure
- 4. 2003.3.28/R.Oka 4 Global feature representation topological geometrical
- 5. 2003.3.28/R.Oka 5 Cellular Feature (Oka,1977), A new cellular automaton structure for macroscopic linear- curved features extraction: Ryuichi Oka, Proc. 4-th International Joint Conference on Pattern Recognition, p.654 (1978). edgeofnorientatioquantized8: edgeofamplitude:),( )),(),,((:),(locationofedge ,0:0,0:1)( )),,((),()0,, 8,...,3,2,1 :conditioninitial ¦¨(i,j) jir jijirji XXX jijirS(i,j ? ??? ??? ? ? ??? ?
- 7. 2003.3.28/R.Oka 7 Location and its Cellular Feature
- 8. Definition of neighbor cells D E C B A 1?? 3?? D E C B A DE C B A D EC B A 5?? 7?? 2?? 4?? 6?? 8?? D A B C D AB C DA B C D A BC
- 9. 2003.3.28/R.Oka 9 Parallel, local and iterative computation for integrating edge features on digital plane Integration algorithm: ),0(),2(),1(),1(),( 2 ),0(),0(),0(),1( :8,6,4,2for ),(),,,( ????? ???? ? ?? DtBtCtAtD t DCAD tDtjiS ??????? ? ??? ? ?