Deliverables 1/16

[bstadt]

Be able to run a subsection of real world data through the pipeline and generate naive statistics about it (i.e. identify & separate clusters, retrieve cluster area, and density of clusters in a given volume)

[bstadt]

Status	Deliverable	Notes
Complete	Finish Naive Cluster Stats Methods	Will
Complete	Threshold Synapse Clusters Based on Size	Will
Complete	Complete Algorithms.md for SLIC and Otsu+ConnectedComp+Threshold	Will
Complete	Finish Writing Tests for the Existing Cluster and Connection Lib Code	Brandon
Complete	Perform Full Pipeline Analysis Testing and Debugging on Real World Data	Brandon
Near Complete (Tuesday Meeting)	Complete Algorithms.md for the PLOS pipeline and NDREG	Brandon
Complete	Generate iPython Notebook investigating cluster volume distributions	Richard
Complete	Pickle histogram data from analysis	Richard
In Progress	Docker algorithms. Docker should take in images and output synapse clusters/locations	Richard
Complete	Incorporate pickle boolean to histogram methods	Richard

https://github.com/NeuroDataDesign/pan-synapse/blob/master/background/PLOS_Algorithms.md.ipynb

https://github.com/NeuroDataDesign/pan-synapse/blob/master/background/NDReg_Algorithms.md.ipynb

https://github.com/NeuroDataDesign/pan-synapse/blob/master/background/connectLib_revised.md.ipynb

https://github.com/NeuroDataDesign/pan-synapse/blob/master/background/SLIC_Revised.md.ipynb

[jovo]

@bstadt

plos synapses

• [x] pseudocode, same deal as #orange-pandas, make sure each mathematical object lives in a very clear space, for all inputs and outputs.
• [x] i did not understand the details of the simulations. please provide proper extensive description of the generative distribution.
• [x] seems like you computed PD, but what about false positives? use ROC curve instead?
• [x] do something for real world data
• [x] comment on algorithms.md in paragraph form about what you like and what you didn't, what wasn't clear.
• [ ] profile code

ndreg

• [ ] finish for next week.
• [x] get dissertation from kwame to read.
• [ ] run on real data rigid and affine

SLIC

• [x] see/implement 3D version

connected

• [ ] upload updated notebook
• [ ] pseudocode
• [ ] run on real data

both: please comment on algorithms.md in a paragraph, pro's, con's, and at least 1 improvement