Boundary Finding
Using Statistical Shape and Texture Models



Intermediate
 

Purpose:  Register important objects in Medical Images to an atlas derived from a statistical model based on shape and texture using the statistical model of the atlas to steer registration.
 
 
 

 

 
 
 
 
Registration Output:  S = B(Sx,Sy,Theta,Tx,Ty,X)

 

 
 
 

Standard Deviation of Weights of Training Set Corpus Callosum


X Y
B1 7.8950 7.5807
B2 4.8196 6.4828
B3 4.4680 6.0843
B5 3.3401 4.5528
B6 2.8939 3.2921
B7 2.6683 3.1939
B8 2.2636 2.7925
B9 2.0434 2.4067
B10 1.8297 1.7894
B11 1.6637 1.7134
B12 1.3479 1.4610
B13 1.2053 1.3249
B14 1.0312 1.1550
B15 .9220 1.10126
B15 .8907 ,8166
B16 .7507 .6517
B17 .6951 .6009
B18 .6333 .5402
B19 .5911 .4776
B20 .5502 .4273
B21 .5267 .3995
B22 .4717 .3399
B23 .4515 .3349
B24 .4225 .3178
B25 .4115 .2945

 
 
 
 

 
 
 

 
 

 
 
 
Shape Value = 136.28
Shape Value = 169.735

 
Shape Value = 73.037
Shape Value = 115.421

 
 
 
 
Shape Value = 39.1354
Shape Value = 43.4352

 



 

Measures of Badness




 
 
 
 
 
 
 
 
 
 
 
 
 
 

Results:
 


 

Enhancements and Refinements