Deliverables for April 24th

[mlee156]

@alee156 & @jonl1096 :

• Before Thursday: Finish Batch implementation and send to Greg for approval
• Thursday: Implement changes Greg suggests after we send the above to Greg

---

---Postponed after talking to Jovo----

• @alee156 & @jonl1096

• By Monday: Make sure web service is functional -> As discussed with Greg, backend code needs to be fixed, will be finished by 27th

@alee156

• I investigated tractography one last time

• Example 1 and Example 2

• As reference: https://github.com/NeuroDataDesign/seelviz/blob/gh-pages/Albert_Jupyter/Final%2BNotebook.ipynb

@jonl1096 Show that bright points function was in fact extracting the brightest points: ipybn

@toekneesunshine

• By Thursday: Look into structure tensor code, noted that last week results looked erroneous. Figure where the bug is. As it turns out, when loading from TIFF --> array we were improperly encoding everything as uint8. This caused underflow problems since small values just got rounded to 0. I fixed the bug and updated tractography.py -- see here.
• By Monday: Discussed with Jovo at office hours on Thursday. Generated centroids plot of each region, with size scaled by amount of bright points within the region. ipynb. Original region plot, centroids region plot.

@lukezhu1

• By Thursday: generate tensor plots using dipy's fvtk.
• By Monday: generate surface mesh of the brain for better viewing. alpha=1 alpha=0.5 s275

[jovo]

committed

@jonl1096 keep the actual 10,000 @jonl1096 demonstrate quantitatively that you got the right answer @toekneesunshine let's get convincing evidence that it is correct? show slides of raw data demonstrating that the regions are reasonable @lukezhu1 update opacity to correspond to brightness of points. make sure she can tell which things are bright. @alee156 try changing that one line of code @alee156 check whether their tracks are similar to your tracks? decide how you will estimate connectivity across regions (by thurs)

[jovo]

@alee156 @lukezhu1 @jonl1096 @toekneesunshine by thurs, please make slides

1. schematic
2. for brightest point stuff
3. and graph estimation stuff (we can improve them later).

ask @gkiar to show you his presentation slides so you know what they should look like.

if you show me a draft before thurs, we can iterate.