Zach Mor Journal

1000000142 hours

Project 1

Python Tutorials

Python Software Page

Project 2

tomviz Software Page

TomoPy Software Page

tomviz Liver Tutorial

TomoPy Liver Tutorial

1/23: hours: 1

• • created journal page

  • explored history of VR

1/24: hours: 1.5

• • exploring development software

1/28: hours: 2.5

• • download paraview

  • looked at data

2/1: hours: 1 

• • some final project ideas:

    • visualizing sound/waves/fluids

2/4: hours: 3

• • download and install unreal

  • begin drafting project plans

  • research sound reconstruction

  • explored data file types 

PROJECT 1 PLAN

Create a guide for bringing VTK files into Python, converting them to NumPy array for processing, and then exporting them out again as VTK files

Activity: learn about relevant CV processes that ParaView operates (slicing, threshold, tinting, cropping, higher level methods), discuss what you cant do in ParaView 

Deliverable: Guide about importing, transforming, processing, exporting VTK files, for example, I will attempt to blur the rat heart vessel. 

[editted and fleshed out here]

2/6: hours: 2

• • read about ultrasound imaging: 2d, 3d reconstruction from sound

  • read about nrrd, .stl, .dcm, .sx file format: to do, add description to wiki contrasting experiences 

2/9: hours: 1

• • explored  researchers at that may have sound data to be visualized

  • ultrasound and fast particle movement were some research interests I discovered 

2/13: hours: 3

• • Research into 3d computer vision

  • drafted project plan

2/19: hours: 5

• • Browsed a bunch of different 3d scene datasets 

  • researched RGB d file type, along with polygonal meshes

  • researching python packages to interface with VTK files 

2/21: hours: 4

• • met with Chris about his research

  • researched xray imaging, reconstruction

  • imported the Rat Heart Vessel into Python

2/23: hours: 3

• • finished processing

    • segmented into four parallel slices

    • added space between them 

    • concatenated to an inverted copy 

2/25: hours: 1.5

• • exported file as a vtu... thought it would be .vtk 

  • failed to visualize on my laptop

  • trouble shooting different ways around the problem

  • added information on wiki about rgbd data 

3/3: hours: 1.5

• • learned more about vtu file

  • still unsure about increase in size

  • scaled down to 30% and actually rendered!

3/4: hours: 2

• • reading more about vtk low level and high level libraries

    • the vtk documentation is AWFUL

  • read a paper related to my project

    • https://conference.scipy.org/proceedings/scipy2015/pdfs/cory_quammen.pdf

  • done with the vtu for now

  • started new process file (material for another guide)

    • goal: normalize and add normalized array to file

3/11: hours: 8

• • began publishing guides in collaboratory 

3/28: hours: 6

• • worked on geometry guide

    • explored visualization libraries mat plot lib and plotly

    • explored the creation of volumes like trivariate normal distributions

3/29: hours: 5

• • finished geometry guide

    • settled on data example

    • redrafted instructions

3/30: hours: 4

• • finished geometry guide

  • began intravalue guide

    • examples and operations

3/31: hours: 2

• • finished intravalue guide 

    • intro, conclusion

    • descriptions

    • redrafted to passive voice

PROJECT 2 PLAN

JOURNAL EVALUATION

• relevance - 4, beginning was for a a project I'm no longer doing 

• documented - 2, been keeping personal messy records, though I ought to edit them and publish on wiki 

• noted time - 5

• pacing - 4, missing a week; caught up with plan/use of python to implement range of densities

4/1: hours: 1000000000

• • finished project 2

  • April fools

4/2: hours: 7

• • Got familiar with Cristoph's dataset (add link here)

  • read about different similar data formats 

    • 360º scans vs Tilt series 

  • read about how tilt series are collected

  • penetrating waves 

4/3: hours: 3

• • learn about computed tomography

    • sinograms

    • inverse radon transform 

4/4: hours: 4

• • researched techniques for accomplishing computed tomography (CT)

    • fourier transforms 

    • back projection

    • algebraic reconstruction technique

    • had missing pieces to aid accuracy: flat and dark field images, used to reduce sensor noise

4/7: hours: 7

• • tried making my own reconstruction algo

    • this was a waste of time. there are labs dedicated to researching how to do this effectively. I should have known to use a module from the beginning 

4/10: hours: 2 

• • researched different libraries

    • decided on TomoPy

4/12: hours: 4

• • had to install with conda 

    • researched what conda is/ what its for

  • downloaded 

  • had lots of problems getting it set up

  • wouldnt find the packages which were reportedly dowloaded

  • messed up path variable most probably

4/13 hours: 5

• • kept finicking with system settings, researching problems

  • somehow got it to run but i guess without fixing the problem

    • it worked once then i had the same problem not being able to import packages

4/14 hours: 5

• • succesfully imported data into python

  • full res

  • discovered ffmpeg 

    • command line tool used to convert video to frames 

4/16 hours: 3 

• • preprocessing

    •  why -log(data)

    • how does finding the rotation center work 

      •  minimize entropy between frames 

    • segfault error  when algorithmically finding the center so i had to estimate

4/20: hours: 8

• •  reconstruction attempts

  • troubleshooting to increase accuracy (reconstructions were 100% noise) 

    • different preprocessing schemes 

    • looked into source code for tomopy methods 

    • other ways to find the center????

4/25: hours: 8

• • stuck :(

  • decided it wont work with tomopy 

  • tried doing their tomopy tutorial but it wouldnt work either

    • dxchange, tooth data set

  • perhaps my computer is not set up properly 

5/1: hours: 7

• • tomopy guide 

    •  ported to colab notebooks but realized its incompatible as far as Conda environments 

5/3: hours: 2

• • write-ups for all the libraries I used and how they can help a python dev trying to work with visualizing data

5/5: hours: 2

• • redrafted most of my documentation

5/10: hours: 10

• • played with tomviz

    • lots of crashing, weird bugs, such as reportedly using 125GB of my RAM (i have 16)

    • misaligned axis meant having to recompute the pipeline multiple times 

    • it worked in the end!!!!!

5/11: hours: 2

• • tomviz activity guide 

5/14: hours: 6

• • tomviz software page

  • python page

  • tomopy tutorial

  • tomopy software page