Global position and comments#14
Conversation
|
Thanks for this Hichem. From the git perspective, it's all good. You've added the reconstructed lambda position which is useful, but the variable we want to add to the CNN is the z position of the hit. Are you making progress on this front? |
|
Hey Nick,
The goal is to add the z position to the panda file then add it after the
convolutional layers. I'm still doing a bit of googling to find out how to
make that work with my setup, I should have something by Monday. Anything
else ?
Best,
Hichem
…On Fri, Jul 31, 2020, 4:28 PM nhaubrich ***@***.***> wrote:
Thanks for this Hichem. From the git perspective, it's all good. You've
added the reconstructed lambda position which is useful, but the variable
we want to add to the CNN is the z position of the hit. Are you making
progress on this front?
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#14 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AMQQYDH23MBWXSTY4EWGICLR6MSO7ANCNFSM4PPZNVCA>
.
|
|
My point is that "LambdaVtxPosz" is not the z-position of the hit. We want to give the CNN information on the incident angle between the particles and the tracker layer. The best way we could do this using "pre-tracking" info is the hit's z position. We should do even better using tracking if we give the hit z and the reconstructed lambda's source vertex position because the lambdas are produced over a spread in z and the incident angle is determined entirely by where the particle was produced and where it interacts with the layer. |
| tree1->Branch("LambdaPt",&LambdaPtKpi); | ||
| tree1->Branch("MCLambdaDeltaR",&MCLambdaDeltaR); | ||
|
|
||
| // Global position vars of rec Lambda added to tree |
There was a problem hiding this comment.
This info is actually already being saved. See
TrackingDstar/LambaAnalyzer/src/LambdaAnalyzer.cc
Lines 1355 to 1363 in 24150c7
No description provided.