perklab / spectroscopy_trackedtissuesensing Goto Github PK

When the module is reloaded, the observer list in self is reset before all the observers are removed

[] Use shapes to differentiate between samples
[] add adjustable axes for easier analysis

TODO
[] Adjust the sampling rate of the spectroscope (This is not currently working in PLUS file for some reason)
[-] Add a counter to consecutively label files

DONE SINCE LAST MEETING
[-] Use seqBrowser to create a sequence of image_image node over a set interval ex 1 second
[-] Get the data collection mode to a functional state. Combining
[-] Change the drop-downs to sliders
[-] General Bug fixes
[-] Add a continuous collection option

DONE
[-] Create the UI
[-] Link the UI to the widget and logic class
[-] Create a seqBrowser
[-] Save the created sequence as a .csv file

• Drop Fiducials/ control points at a specified location
• Change colour of dropped fiducial depending on the classification output
• Transform control points to new coordinate systems
• Drop Control points on tip of probe in the world reference frame

[x] Divide my spectra by the characterisitic curve of the light source
[x] Alter my useable wavelength to where the signal to noise ratio is acceptable. The signal explodes for areas where low light is coming from the source
[ ] Add the functionality to the process library
[ ] Apply this to the visualized spectra in 3D slicer

• Create the selector
• Extract the file path and save it to the parameter node whenever it changes
• Use it to load in the model
• Save the path history
• Load the last one by default

Separate my 881 pipeline into formating, preprocessing, training, testing, and associated functions scripts

• Setup the EM tracking system to work in my module
• Get the live transform and place points at the tip
• synchronize a needle model with my probe

There are a variety of best practices which Tamas is teaching me.

• Don't use self.variable to store data as this is disconnected from the scene
• Using VTKObservationMixin

• Change the if statements to be more robust
• Remove commented-out items
• Rename functions

• Revise the pipeline to use the slicer data
• Remove the testing portion of the training
• Save the trained model
• Load in the trained model to slicer
• Classify the spectra live in 3D Slicer

[-] Write the SOP
[-] Run through a trial of the SOP with Amoon

[ ] Module: Implement the changes he laid out in our meeting
[ ] SOP: Write a new draft of the SOP and send it to him for review

[ ] Make it so the save button disables while data is being saved before re-enabling

Instead of checking if a node type exists, check if the node ROLE is present in the parameter node. If it is not then save an empty node of that type to the parameter node

[ ] On line 125 in onConnectButton Clicked:
I should instead check if ROLE: connector is saved to the parameter node. If there is nothing with that role then I know to create one and save the ID to the parameter node.