Spring 2012:Vincent Lee Journal
- 1 Week 1 (2/6-2/12)
- 2 Week 2 (2/13-2/19)
- 3 Week 3 (2/20-2/26)
- 4 Week 4 (2/27-3/4)
- 5 Week 5 (3/5-3/11)
- 6 Week 6 (3/12-3/18)
- 7 Week 7 (3/19-3/25)
- 8 Week 8 (3/26-4/1)
- 9 Week 9 (4/2-4/8)
- 10 Week 10 (4/9-4/15)
- 11 Week 11 (4/16-4/22)
- 12 Week 12 (4/23-4/29)
- 13 Week 13 (4/30-5/6)
- 14 Week 14 (5/7-5/13)
- 15 Week 15 (5/14-5/17)
Week 1 (2/6-2/12)
- What do you want to get out of 20.345?
I want to make awesome optical things and learn about optics. Also interested in seeing how this all applies to bio, but more interested in the optics.
- What are your areas of interest?
I am interested in looking into the physical optical side of things, less the image processing. We'll see.
Week 2 (2/13-2/19)
Progress: Working on optical trap lab.
Week 3 (2/20-2/26)
Week 4 (2/27-3/4)
Week 5 (3/5-3/11)
Week 6 (3/12-3/18)
Progress: Continued building fluorescence microscope. Having some trouble getting alignment correct, but getting there.
Week 7 (3/19-3/25)
- What is the most important thing you accomplished during the first week of your project?
We primarily decided on a design and what materials were required. We ordered these materials, and will bring the design together when they arrive.
- What three things do you plan to accomplish in the next week?
Get materials and start building
- What three things do you plan to accomplish in the next month?
Get lightfield data from the camera and process it into a focusing image.
Week 8 (3/26-4/1)
- Did you enjoy Spring Break?
Week 9 (4/2-4/8)
- What is the riskiest element of your project? What is your backup plan if that element fails?
Riskiest is really the processing. If we cannot successfully do it by hand, we will need to fall back to the Lytro software, which may or may not result in lost functionality.
- List two skills you will have to learn in order to complete your project.
Lightfield processing in MATLAB Retrofitting existing optics into a microscope layout
- What companies make products related to your project? Whose products are the best?
Lytro is basically the only company that makes products at all related to this, but these products are not for microscopy.
Week 10 (4/9-4/15)
- List at least five resources that have helped you. Your list may include papers, documentation,, textbooks, videos, downloaded code, or people who are not members of the teaching staff.
Recording and controlling the 4D light field in a microscope - Levoy, M. Code & Article from <http://eclecti.cc/computervision/reverse-engineering-the-lytro-lfp-file-format> Code from <https://github.com/nrpatel/lfptools> Software from <http://graphics.stanford.edu/software/LFDisplay/> http://graphics.stanford.edu/software/LFDisplay/lfmintro/
- What was the most fun part of your project so far?
- What task are you procrastinating the most?
Week 11 (4/16-4/22)
- List at least three statistical techniques you plan to use.
- Generate simulated data sets and test these techniques.
- What was the part of the project that you originally didn't think was important, but now has become a major task?
I really do not see any part of the project following this path.
Week 12 (4/23-4/29)
- Is there any part of your project that you now realize is not so important, and should have been skipped?
To a point, trying to understand the Lytro so much was not needed, since really only the minimum focus was needed to get all of the information needed.
- What "unwritten knowledge" have you found out in the process of conducting your project (e.g., the kind of thing that never gets described in a book or paper, but you have to experience to understand)?
- Design a mathematical model, a simulation, or small pilot project, which takes no longer than 3 days, to reduce the risk of the remaining part of your project.
Take images of two airforce targets stacked and crossed. At least one will be off of center focus. Digitally refocus to second target.
Week 13 (4/30-5/6)
This week, I have been trying to modify LFDisplay to work with hexagonal grids, but the code is difficult to understand and I suspect/fear that the lightfield processing happens in an OoenGL shader. I am, today, switching my focus to writing a more minimal program that processes the image in the LF -> 4D Freq Space -> 2D Freq Space -> 2D Spatial method laid out by Ng in .
5/4: So far, have written a way for the program to import an image, pre-process it, and properly select out lenslets. Need to read the Fourier Slice paper in greater detail in order to know how to best do the 4D Fourier analysis and how to correctly filter the image. 5/4: Reading paper. Is dense. 5/5: Still reading paper, but I have a good idea of what is going on now.
- What is the biggest thing you would have done differently in your project, given what you've learned throughout your project so far?
I would have started worrying about software a lot earlier. Software is really posing the largest problem here.
- Instruments often require a biological demonstration to make their value clear. What would the most compelling demonstration of the power of your technology be? What will you do during the class period?
Golgi apparati would be nice. Basically, any depthy sample will prove interesting, and we should optimize for transparency. We will also be taking data from fluorescent beads, possibly suspended in gel. The power of the technology lies in being able to refocus and basically do confocal post-process. There is also a possibility that the technology lends itself to being able to see behind obfuscating parts of the image, as seen in more photography-based applications of seeing behind bushes, etc.
- Reflect upon your pilot project from the previous week: did you carry it out? Why or why not?
Did not yet, because software is still being worked on. DID take all of the necessary samples.
Week 14 (5/7-5/13)
- Have a function that allows us to go from the Lytro to a lightfield processor.
- Depthy, macro images are good, but micro images have problems with depth. Need a higher M.
- However, need a larger illumination source. All depth data transforms into moving the illumination source around, which is not cool. Beam expand the shit out of the LED?
- If you had twice the time to do the project, what would you differently? What about if you had half the time? How does this thinking change your actual plans?
I would have taken the Lytro apart right away. The development suffered from problems with the optics that disappeared when the optics disappeared. Now, this just means work as hard as possible now that the optics are off.
- What have you noticed about your own mind, in the way that you create ideas, implement ideas, and test ideas? Consider both cognitive and emotional variables.
I find that I tend to come up with a lot of ideas that may be a little far off, and know how to test these ideas and possibly even implement them, but implementing them is hard. Some things I feel very motivated to get done, usually when they play to my comfort domains. Some things get pushed to the back burner.
- What was the most important resource you now use, that you did not know about at the beginning of the term? How can you find more resources, more rapidly, in the future?
Matlab help. It has been gloriously helpful. I am pretty sure how I can and how I will find more resources in future is through searching on the internet.
Week 15 (5/14-5/17)
- What have you learned about working with others? Are there proactive or responsive things you should do differently, to maximize work with others?
I think the thing that got in the way most this term was time allocation. Communicating scheduling might have helped, although I feel it would have helped mostly from a motivation/obligation standpoint.
- If you had twice the resources to do the project, what would you differently? What about if you had half the resources? How does this thinking change your actual plans?
I am not sure if it would have changed the project. We consciously chose to use a cheaper alternative (Lytro) to see what we could get out of it. We had the opportunity to use an expensive system, but opted out.
- Based upon your reflections of the previous week: are there things you can practice, to make you more efficient at creating, implementing, and testing ideas?
The most important thing is motivating myself to get stuff done and out of the way. Energy drinks and no regrets?