Home » Blogs » bouchet's blog

Comparison KFParticle vs. Helix swimming for secondary vertex reconstruction

Updated on Fri, 2015-04-03 08:45. Originally created by bouchet on 2015-04-03 08:08.

Simulation :

  • Y2014a geometry
  • 20k events ; 1 D0 per event ; 0.1 < Pt < 5 , flat 

KFParticle :

Helix swimming :

--> Both methods are compared with the GEANT decay
--> use .fz file (g2t_vertex table) to get the position of the D0 decay)

A) X,Y,Z componants of secondary vertex
1) KF vs. GEANT

2) HS vs. GEANT

3) KF-GEANT vs GEANT

4) HS - GEANT vs GEANT

5) Comparison of the mean and sigma for both methods 


Comments :

  1. no apparent shift vs GEANT for both methods
  2. resolution for KF is slightly better (~10microns) than HS

B) Decay length plot (3D)
Plots are :
  1. X-axis : decay length from GEANT (3D)
  2. Y-axis : decay length from KF/HS (3D)
  3. marker : sigma from the FitSlicesY()
1) KF -GEANT vs GEANT


2) HS -GEANT vs GEANT

same but zoomed :


comments :
  1. the resolution of KF for the decay length is ~60 microns, flat ; 
  2. there is no optimization of KFParticle, ie cuts on probability, etc ...
  3. the resolution of HS for the decay length is ~ 50-55 microns but it seems also that HS fails for the low decay length ; the difference seems not to be flat for low decay length (< 200 microns)


 

»