But now the new website is done! It’s now at http://www.asymptoticdesign.org/, and I’ll be updating stuff there for now on — so be sure to check there often

All of the current blog material has been moved over to the new site, and will remain here for the foreseeable future.

I’ve always been surprised at how expensive professional microscope imaging software is, and how *repetitive* the tasks are for it. It’s almost like photoshop… you just need a ‘count number of particles’ button, a ‘calculate intensity’ button, ‘calculate area’, and so on… So I figured I’d try my hand at writing some basic image analysis software and see where it takes me.

The result is ‘OpenScope Software’, which I intend to release as a software package to go with the hardware for OpenScope. Right now there’s two tasks I want to be able to do with fluorescence microscopy:
-Calculate the intensity/brightness of a selected region of interest
-Track that region of interest across a sequence of images so I can track the fluorescence vs. another parameter (such as optical sectioning, introduction of a drug, etc.).

OpenScope software uses Python Imaging Library and TKinter to do just that. It creates a RegionOfInterest object that track a region of interest, can draw a box around it, and computes values using the pixel data (including when the pixel data changes). It also has a Sequence object that handles loading a series of images labelled numerically (such as image01.jpg, image02.jpg, image03.jpg… until the last image) so that you can perform image analysis on optical sections or videos of fluorescence microscopy.

The code is available on github here: https://github.com/asymptoticdesign/openscope-software; the README file includes the instuctions for the few features it has, along with a test image.

Coming Soon: Some more software features and a physical fluorescence microscope to go with it!

The class I’m teaching is focused on teaching how to make generative and visual art in Processing; the class starts with basic concepts in programming (like control flow) and Processing (like colors and primitive shapes), gradually adding concepts like randomness and complexity to really make Processing shine. The current notes (and lots of programming exercises!) for my class are available in their own github repo. I’ll be adding materials there for the next two weeks until the class is over; by then the entire course should be available online!

And now its time to show some neat processing sketches I’ve been working on! This first one was inspired by the work of Rachel Boyce, who did a really neat sketch utilizing overlapping spirals. I was so impressed with her work that I decided to modify it myself! I called this piece ‘Moire Spiral’ because the overlapping lines give yet another moire-type pattern (you can tell that I am an enormous fan of these patterns!).

The next piece is called ‘Search Light’, because it is reminiscent of a searchlight searching for something in the night. It was actually generated from code that was meant to make a ball travel in a circle; I next made the ball move in a circle modulated by Perlin noise, and then finally drew a line from the origin (top-left corner) to the ball. If you look closely at the tip of the ‘searchlight’ you’ll notice that its still moving in approximately a circle!

This last one was a further modification of ‘Search Light’; instead of having one point fixed at the origin, I had both points of the lines moving in overlapping circled overlapped by noise. I really like the aesthetic effect here — it almost looks as if there is a long ribbon turning and rippling through space. Check it out!

I’m hoping to write more about my microscope this weekend; it’s coming along quite well and now it’s a full-fledged fluorescence microscope! Hopefully I’ll get some sweet pictures of the images it can turn out in addition to some more schematics.

Conceptually, a bright field microscope needs surprisingly few specialized parts (a bright field microscope is what most people think of when they think of a microscope — you see an object backlit by the microscope light). Most of the magic happens in the objective — and those are readily available from the optics supplies like Thorlabs or Edmund Optics. Besides the objective, you’ll need a couple of convex planar lenses (not too difficult to come by) to focus the beam, and a mirror to help with alignment.

A simplified schematic of the optical path is below:

A red LED is used to provide the brightfield light; technically a blue LED would be better (because blue is a smaller wavelength and therefore provides a higher resolution), but would make things a bit messy when I integrate fluorescence later on. The LED light hits the sample and scatters from it; the objective collects the light and needs a 200 mm lens (f1) to correctly collimate the light. The collimated light gets directed to the middle of the camera’s field of view with the mirror, and voila, you have a microscope!

Here’s an image of my constructed microscope (made from mostly Thorlabs’ parts):

I tested it using a standard with line thicknesses of known size:

and could detect the smallest lines (~18 lines/mm) with pretty good resolution:

I also used a target with 600 lines per mm (meaning the lines should be around 1.6 um apart), and was able to resolve the lines. So we’ve got a resolution of around a single micron on this microscope — enough to make microspheres and most cells visible as blobs.

Right now I’m working on adding a green laser to the optical path for fluorescence microscopy, and figuring out a way to reduce the cost of the microscope. It looks like the final fluorescence microscope will cost in the $1000-$2000 range (in large part because of the camera); I’d like to drop that cost down by about $1000 ideally to the $500-$1000 range. If anyone knows of a good place to get cheap (but not crappy) optics, computer vision cameras, etc., let me know!

Let me review their primary arguments:

It requires that final manuscripts of researchers’ journal articles that explain, interpret and extensively report the results of federally-funded research — manuscripts which have undergone publishers’ validation, digital enhancement, production, interoperability and distribution processes — be publicly available online, worldwide, no more than six months after publication.

Let’s view the text of the bill, shall we?

SEC. 4. FEDERAL RESEARCH PUBLIC ACCESS POLICY.
(b) Content- Each Federal research public access policy shall provide for–
(1) submission to the Federal agency of an electronic version of the author’s final manuscript of original research papers that have been accepted for publication in peer-reviewed journals and result from research supported, in whole or in part, from funding by the Federal Government;
(2) the incorporation of all changes resulting from the peer review publication process in the manuscript described under paragraph (1);
(3) the replacement of the final manuscript with the final published version if–
(A) the publisher consents to the replacement; and
(B) the goals of the Federal agency for functionality and interoperability are retained;

Seems to me that this bill wants the final peer-reviewed paper submitted to the journal, but does not require the final published manuscript with the publisher’s “validation, digital enhancement, production, interoperability, and distribution process” (does anyone else want to know what the heck those words even /mean/?). The alliance’s next point:

The one-size-fits-all six-month deadline for every federal agency that funds research ignores well-known significant differences in how each research discipline discovers and uses individual articles, periods that can last several years before costs are recovered.

First off, the bill is not one-size-fits-all. To visit the text again:

SEC. 4. FEDERAL RESEARCH PUBLIC ACCESS POLICY.
(6) long-term preservation of, and free public access to, published research findings–
(A) in a stable digital repository maintained by the Federal agency; or
(B) if consistent with the purposes of the Federal agency, in any repository meeting conditions determined favorable by the Federal agency, including free public access, interoperability, and long-term preservation.

So each federal agency is allowed to come up with their own repository, as long as they are making a good-faith effort to follow the spirit of the bill. Back to the AAP’s letter:

It limits where government-funded researchers may publish their work.

I can’t find any text in the bill to counter this with because the bill simply does not say this. The only thing that could possibly limit where an author could publish their paper is the publisher’s copyright policy. This act does absolutely nothing to change that. Let’s see what the AAP has to say next:

It undermines publishers’ investments in new business models that currently provide unprecedented access for the public to such works for free or at modest cost.

You couldn’t possibly be referring to Springer’s Open Choice publishing policy, that offers open access for the modest cost of $3000.00 for the author, could you? Or perhaps you mean Elsevier’s Open Access Policy, which charges authors $3000 a pop? How generous of you! And the AAP’s final point:

At a time when Congress is looking to cut unnecessary expenses in federal government and focus budgets on priorities, FRPAA imposes additional costs on all federal agencies by requiring them to divert critical research funding to the creation and management of new databases, archives and infrastructure to handle dissemination of these articles — functions already being performed by private-sector publishers.

Are you serious? The (direct) maintenance cost of arXiv is a tad under $600,000 a year. According to the NSF’s Funding Profile, they awarded 8,509 grants with an average grant size of $165,000. That means that an open access archive would cost as much about about 5 research grants. If multiple federal agencies shared a single archive, the cost could be amortized to less than one grant a year.

As an additional note: the FRPAA only applies to federal agencies with extramural research expenditures (meaning: they give out grants summing to more than) $100,000,000. One hundred million dollars. That means that the arXiv would cost an agency about 0.6% of its external research budget. What do you think their internal budget looks like? I bet it’s a drop in the bucket.

After reading through the above letter, I finally mailed out letters to all three of my representatives — Senators Scott Brown and John Kerry, and Congressman Michael Capuano — asking them to co-sponsor and/or otherwise support the Federal Research Public Access Act. The contents of my letter is below:

Dear [Representative]:

I am writing to ask you to co-sponsor the Federal Research Public Access Act of 2012 ([S.2096/H.R.4004). This bill provides an important mechanism for ensuring that American taxpayers have free open access of peer-reviewed scientific articles funded by U.S. government agencies. It continues a growing trend by both federal agencies and higher education institutions to prioritize open access of research results

One of the fundamental goals of science is not only to perform research, but to disseminate the results to the rest of the scientific community. Yet the cost of journal subscriptions and witholding of research results prevents knowledge from being disseminated as widely as possible; this stunts the rate of scientific discovery by forcing scientists to constantly recreate each others’ work. As the old saying goes, “Six months in lab can save you an afternoon in the library” – without access to current research, countless hours of research efforts are going to waste.

As both a student and researcher at the Massachusetts Institute of Technology, I have access to all of the scientific journals that my institution subscribes to. I use them nearly every day in my professional life; however, I also have access to these manuscripts for projects unrelated to my professional duties. Significant work that has helped others has come out of my efforts as an entreupeneur and citizen-scientist – and much of it would not have been possible without access to up-to-date scientific research. As both a potential author and subscriber to research affected by this bill, I enthusiastically support it.

Not only does the FRPAA provide open access, but it is flexible with regards to the interests of journal publishers and federal agencies. The bill asks agencies to develop their own individual general guidelines, recognizing that there is no ‘one-size-fits-all’ solution. The bill provides an exclusive six month window for journals to provide access to the published material, and only peer-reviewed manuscripts (not the final published edition) are required to be made available.

Please support and consider co-sponsoring the Federal Research Public Access Act of 2012.

Thank you for your time,

Nathan S. Lachenmyer

My position as a researcher (at MIT, no less) puts me in the minority of people that have access to most peer-reviewed publications out there. I’ve used that wealth of knowledge to learn more about current research, help me find solutions to difficult problems in classes and personal projects, and even in my recent work at H@cking Medicine (where I used sensors to come up with a quantitative way to measure symptoms of patients diagnosed with Parkinson’s Disease). A large portion of this work would have never happened without access to current publications; and for that reason alone I want to enable others to have that same experience.

I encourage you to send a letter to your representatives! Here’s how:
-Find out who your representatives are via services at the websites of the House and Senate.
-Write a letter to your representatives! Feel free to use reuse parts of my letter above. The Alliance for Taxpayer Access has some great talking points as well.
-If you’re feeling like being extra active, please support the sponsors for the bills — a list of the sponsors in the Senate and House are readily available.

In the last month or so, the issue of Open Access Research has really taken center stage in the scientific community. A boycott against Elsevier (one of the three largest scientific publishers) was initiated, which snowballed into increased awareness of the Research Works Act. This wild month of attention came to a surprising turning point this week: Elsevier officially withdrew its support for the RWA, effectively killing the bill.

For those of you who aren’t aware of the significance of what I’m talking about, let me provide some background information:

The Problem

Traditionally, research journals have existed to assist in the dissemination of (scientific) information. Academic journals have existed since the mid-1600s, where the job of accumulating work by scientists, typesetting it, printing the journals, and distributing them was a non-trivial task that required a great deal of man-hours and resources. During the early days of journals, they served an invaluable role is the process of disseminating scientific knowledge to researchers around the world.

Typically the editorial board of such a journal consists of distinguished scientists (usually university professors) from a specific field (e.g. a nuclear physics) that manage the process of selecting articles for publishing. These editorial board members are usually volunteers; their role on the board of a journal is considered to be both an honor and a part of their scholarly duties. The editorial board typically matches submissions with referees, whose job is to evaluate the paper’s scientific merit (these referees are also volunteers, and often university professors/researchers). If the referee and editorial board decide to publish a submitted paper, the author (also typically a university professor or graduate student) is then required to sign over the copyright for their work to publisher, and receives no financial compensation for their work — after all, they chose to be scientists to do research that will further humanity’s body of knowledge and hopefully improve our lives. Sharing of results is an integral part of research.

So where do these papers go? Well, many of those some researchers who participated in submission process. They are either directly subscribing to these journals, or more frequently, their university has a subscription to these journals.

Before the advent of the internet and electronic typesetting (like LaTeX), the journals took on a very laborous and difficult job. But now information is primarily disseminated over the internet; authors typeset their own papers; and the cost of publishing (both in print and online) has dropped dramatically.

And this is where the problem begins — even given how much easier it is to publish and distribute information these days, journals frequently cost thousands of dollars per title per year. For example, Elsevier-published journals cost an average of $6,598 a year. Elsevier’s profits, on the other hand, are upwards of $1,000,000,000.00 annually. Yes, that is one billion dollars. Not bad for a company that is outsourcing most of their work to academic volunteers! Plus, they get to keep the copyright to the journals to boot — often keeping extremely strict rules about where and how authors are allowed to distribute their own papers, meaning that most scientists cannot submit a paper to be published in a journal *and* make it freely available to the scientific community.

So why does the scientific community continue to participate in this system? Prestige. Journals are essentially ranked — certain journals have a reputation of having higher standards or publishing more high-impact work than others; as a result, a publication in a high-ranking journal like Nature can make a researcher’s career (resulting in near-guaranteed admission to graduate school, appointment to a professorship, or even tenure). Without an overhaul of the entire system, and of the scientific community’s perceived value of publishing in reputable journals (often published by companies like Elsevier), many researchers feel that they must ‘publish or perish’.

Current Events

Taking advantage of the aforementioned benefits of electronic dissemination, open access archives of scientific papers such as ArXiv and the Public Library of Science were created to serve as centralized databases of published researcher. These databases got around journal copyrights by serving as databases for ‘pre-print’ papers — essentially rough drafts of papers that were available before the ‘final’ paper was published. In addition to these open access archives, the National Institute of Health formed PubMed, an open-access archive that required all recipients of NIH research grants to submit a copy of their final paper under the belief that publicly-funded research should be publicly-accessible.

However, in December 2011 the Research Works Act was introduced, which would effectively nullify the efforts of the NIH and PubMed. The act prohibits any sort of open access mandate for federally funded work — including the National Institute of Health, National Science Foundation, and Department of Energy (all organizations that offer large grants for scientific research).

The Research Works Act (RWA) caused quite a stir in the scientific community; providing a spark that pushed several researches into action. The Cost of Knowledge was started to spread awareness of the RWA and begin a boycott of all Elsevier-published journals. Amazingly, over 7000 researchers have signed it — refusing to publish in any of Elsevier’s journals, some of which are the aforementioned prestigious journals that can make a career. I am proud to have added my own name to the list — I would love nothing more than to see an end to these ‘closed-access’ publishers that take advantage of the scientific community for large profits.

And this brings the story to the present. It seems that all of the RWA-related noise had an impact — Elsevier withdrew its support for the RWA, and the representatives that originally backed the bill dropped support for it too. For now, it appears that the RWA is dead.

Not only is the RWA dead, but there is now a bill being pushed through Congress — the Federal Research Public Access Act — would require that research funded by eleven federal agencies (see the wikipedia page for a full list) be open access to the public, in a searchable archive similar to arXiv or PubMed. The passing of this bill would be a huge success for open access, and would make a huge portion of new scientific research (a lot of research gets funded through these federal agencies) available to the public, free of charge.

Take Action

There’s lots that you can do if you’re interested in supporting the FRPAA and want to help end the absurd policies of academic journal publishers. Things that anyone can do:

-Read about the FRPAA; learn about it, and form your own opinion on it!
-Sign a white house poetition to have President Obama review the FRPAA if you want to support it.
-Write a letter to your local representatives asking them to co-sponsor and support the FRPAA!
-Spread the word! After all, we’re looking to improve dissemination of information — tell your friends, discuss it with them, form opinions, and take action!

If you’re working in academia, additional things you can do:
-Refrain from publishing in an Elsevier journal — you can read more (and sign a list of scholars that are boycotting Elsevier) at The Cost of Knowledge.
-John Baez at Azimuth made some sweet posters to place on your office door to raise awareness about the Elsevier Boycott. Print one out and post it on your door!

On 2012 Feb 15, I received an e-mail about an event called H@cking Medicine. The e-mail said that it was a conference aiming to bring “Engineers, scientists, physicians, and entrepreneurs, in one location, creating disruptive healthcare solutions today.” An opportunity to create disruptive tehchnology piqued my curiosity, and I decided to at least read more about it.

Reading about the Fall 2011 event, I found that it focused on ‘Big Data’ (Aquisition and Analysis), ‘Biosensing and Biostimulation’, ‘Healthcare Automation’, and ‘Synthetic Biology’. Now, those were some topics I could get interested in! I was sold on the event, and decided to apply to be one of 80 participants. A few days later, I received a response that said my application was been accepted, and I was awarded a scholarship to cover the conference fee.

2012 Feb 25 9:30 — I arrived at the conference. Checked in and got my name tag, and was introduced to the first quirk of the event — I had to pick a colored sticker to label what my background was. The options were ‘Engineer’ (Orange), ‘Business’ (Blue), ‘Healthcare Professional’ (Green), or ‘Design’ (Pink). As much as I like to think of myself as a creative design type, I grabbed an orange sticker and put it on my name tag. I grabbed a sharpie and clarified: EECS/BIO. I wouldn’t want anyone to think I was one of those software types.

The room was already full of about 50 people (of the expected 80), many of them already chatting and forming groups. Turns out the little stickers were pretty unnecessary — I could tell everyone’s background just by looking around the room. Nice collared shirts, slacks, shiny shoes — the venture capitalists and entrepreneurs. Software and web developers wore jeans and tight black t-shirts. Anyone with a background in design had a business-casual collared sweater on. And not a single other pair of BDUs in the entire room. Sigh.

Socialization was not going to happen until coffee did, I realized. So I took care of that first, and with a donut in one hand and some go-juice in the other, I found grabbed an open seat at a table. Rejuvenated by caffeine, I found it in me to start chatting with the people at the table. Three out of four were physicians, and it was really eye opening to hear their views about technology in healthcare. The bottom line I got from several physicians:

-While you’re a resident, your focus is in making sure that people don’t die. Anything that isn’t directly related to patient care (like learning new technologies) is considered a waste of time.
-Most of the techniques that are taught and shared focus on time-saving, not accuracy or efficacy.
-Seven years later you become a fellow, and you haven’t done any real experiments, science, or academic learning for most of your residency. You’re so far behind that you don’t have much time to be able to pick up new things that aren’t directly related to your career goals.

Frequent words thrown around (by healthcare professionals!) about their technology: ‘archaic’, ‘broken’, ‘technology-adverse’, and at times, ‘barbaric’! My impression was that there is definitely work to be done in healthcare — there’s tons of problems that need to be solved; now they just need people willing and able to solve them.

2012 Feb 25 10:30 — The event starts. The organizers say a few brief words to get everyone pumped, and Jason Jacobs kicks the event off with the story of his own startup (RunKeeper). I’m not usually one for speakers kicking things off with inspirational stories, but Jason had some very practical advice to give about solving problems and starting a company

-Don’t overanalyze, just do something. You can analyze it when you’re done with it.
-Keep a broader vision in mind for the big pictures, but work on narrow, concrete problems that you can ‘hit out of the park’
-work with people that you get along with and are likeminded rather than the most skilled.

Very wise words!

2012 Feb 25 11:00 — The event starts to really get going when the organizers ask for some problems that need to get solved. Ideas start flying from crowd, ranging from conservative to extremely radical, and from the specific to broad and general:

-Cost transparency in hospitals
-leveraging social media to encourage healthy lifestyles
-public-access health data
-improve inefficiencies in healthcare (scheduling, automation, etc.)
-centralization of health information (especially about medication side effects, alternatives, etc.)

And then they started asking for solutions, and even more creative, radical, amazing ideas started coming out from the crowd. We could’ve been there for hours just listening to ideas, but when it was obvious that everyone was ready to get hacking, they let us loose for a while to intermingle and start meeting people.

So I wandered from group to group, trying to figure out how I could pitch myself to a project. I had EECS/BIO and HARDWARE/DEVICE DESIGN scribbled on my name tag, and was hoping to find someone that could use some hardware design in their project. As I schmoozed with physicians and entrepreneurs, it became obvious to me what their problems were — they wanted bigger, better data. Data that was continuous instead of discrete; data that was convenient to access, data that could give them quantifiable results and diagnosis rather than subjective interpretations. So I ventured from group to group, pitching my services as a ‘data accumulator’ — tell me what data you want, and I’ll find a way to accumulate it.

It was at this point that I met Verena, Wael, and Sean. Verena did neuroimaging, Wael was a neurosurgeon, and Sean was a neuroscience PhD at BU. Chatting with them, they were discussing how neurological disorders are so difficult to treat because of how subjective most tests are. This means that patients often get mixed messages from different physicians, and even the same physician has trouble tracking a patient’s performance over time. I suggested using sensors to measure symptoms, and perhaps even having a mobile sensor that a patient could wear that could provide a continuous stream of sensor data (while the patient is using it). We were all liking the sound of this idea; we knew that we had a team.

The organizers called everyone back, and we did one more round of problems-solutions, with everyone keeping careful track of who was proposing what idea. Everyone was already storming to form their teams, and decide what they wanted to work on. After all of the ideas were proposed, we were off to the races. Everyone grabbed some team members and started hacking.

25 Feb 2012 15:00 — My team has spent about 3 hours straight just figuring out what the heck we’re going to do. We want to do quantitative measurements of symptoms to help physicians make better decisions. Ideally we want the device to be something that a doctor can send with the patient, or utilize items that the patient might already have (consumer electronics) so that the doctor can get data from the patient between visits. We brainstorm, and decide that Parkinson’s is a good place to start — there’s a huge number of people affected by it, and the symptoms (tremors/dyskinesia) are easy to detect with hardware that’s reasonably ubiquitous (accelerometers). So what if we made an accelerometer-based device that could store the information for the doctor? That’s brilliant!

Unfortunately, it’s been done. Crap.

What about using an iPad’s touch screen to measure tremors? We could have patients trace letters and measure their variations, and correlate that to their prescription — oh, wait, that’s been done too.

(Google Patent is your friend! “Six months in the lab will save you a day in the library”, so they say.)

Time is flying by, and while we have lots of ideas, we don’t have a concise product to make, an audience to sell it to, or why its different or better than what already exists. But at least we have a name — ParkinSync!

25 Feb 2012 18:00 — Still no concrete plan. The organizers call us back to reconvene, and share our initial concepts with the rest of the teams. Lots of great ideas; though it seems like no one has really made much progress either. For some teams, tonight will be a long, long night.

But not for us. We decide to reconvene in the morning the figure out our plan from there.

26 Feb 2012 9:45 — I arrived a bit later than I did the day before. The conference room seems emptier — maybe some teams got discouraged and decided not to continue on with their project? Again, time for go-juice and donuts. No way is my day going anywhere without caffeine.

We learn that the proposals are due at 15:30 today. That gives us about 6 hours to wrap up our project. Time to get to work.

26 Feb 2012 10:15 — Our group is entirely reassembled and we start coming up with ideas again. We think we’ve finally got an idea that’s worth pitching — make a mobile application that uses the phone’s accelerometer to measure the patient’s state. The doctor sends data to the patient to remind them to take their medication, they take a quick diagnostic test before they medicate, and the results get sent straight back to the doctor. After a few days of data collection, the doctor can review the data and send back recommendations to the patient — without having to schedule a visit.

The idea is great. It’s concise, narrow in scope, but easily extendable. And now we need a prototype and a business plan. We split up the group into two sides — the physicians (with their much more intimate knowledge of the field) come up with a marketing plan, and the software/hardware hackers start figuring out how we’re going to get this system working.

26 Feb 12:00 — Realization: Our product is a mobile app. None of us are familiar with the Android SDK or iOS. Crap.

So we start reading about the Android SDK and figuring out how we’re going to get an app that works.

26 Feb 2012 12:30 — Realize that learning Android by 15:30 is not going to happen. Mock up some pictures of what the app would look like and call it done.

Realization #2: Our product needs to send data back to the doctor somehow. The doctor is probably going to review the data through a web interface. We don’t have a web developer on the team.

Time to start reading through Django tutorials and see if we can set a database up!

26 Feb 2012 13:30 — Still no progress on Django. On the bright side, Django looks awesome and now I want to learn it! Unfortunately, it’s not going to happen by our deadline.

Solution: Mock up the website in Adobe Illustrator.

Wael, Verena, and Michael are making serious progress on our business plan and slides. But we still want something to show off as our prototype. I may not be a mobile app or web developer, but I sure as hell know how to use an Arduino and Python. Time to get to work!

26 Feb 2012 14:00 — Went back to my office, grabbed an arduino and an accelerometer (that I cleverly grabbed the night before in case I needed it), and got to work. Acquiring data of unsteady hand motion wasn’t too hard; it was actually data analysis and interpretation that was more difficult.

26 Feb 2012 15:15 — Success and making some pretty pictures of accelerometer data! We have a demo to show! The rest of the team is putting the last touches on our presentation, and we ship it out just in time at 15:30.

26 Feb 2012 16:00 — Presentations begin! There were around 15 teams at the end of the competition. Due to the large number of groups, we do 5 minute presentations with about a minute to do Q&A with the judges. Lots of great ideas, and lots of progress made — ranging from android apps meant to help you share information about and decide on a PCP; scheduling software that utilizes machine learning of patient habits to optimize the doctor’s schedule and minimize double-booking;

26 Feb 18:00 — Winners are announced!

Best Presentation: MyBetterFit
Most Disruptive Technology: ParkinSync!
Most Progress: Universal Prosthetics

Holy crap! We won Most Dispruptive Technology (and with it, a $1000 check for our team)! All of us are honestly surprised — we thought we had a solid idea, but didn’t think that we had developed it to the point of being a competitive idea. But apparently the judges thought we had a fantastic idea! After the awards were wrapped up, the team got together and started talking about our victory — and now phrases like ‘provisional patent’, ‘intellectual property’, and ‘clinical validation’ are being thrown around left and right. Are we going to start a company from this idea? What’s the next step?

We’re all pretty overwhelmed at this point, so we decide to meet in two weeks or so and talk about it once we’ve had a chance for the adrenaline to wear off. And that’s where we stand — with a great idea, and potentially taking the first steps to turning our idea into a marketable product.

I wish I could’ve shared more details about our project, but I’m going to be filing a provisional patent — I’ll be sure to talk about it once I get the patent done!

The exhibit consisted of Sandtrap; and Processing In Digito, two pieces that I think explored different aspects of art that I enjoy.

Sandtrap focuses on interaction and creativity, relying on participants to really drive the piece. Without anyone to participate, there isn’t really much to Sandtrap. Processing In Digito, on the other hand, is a piece that is conceptually diametric to Sandtrap. It is computationally generated — computer algorithms produce the imagery, and the artist is essentially absent from the piece. In a sense, Sandtrap focuses on agency while Processing in Digito focuses on removing it.

Below are some pictures of Processing In Digito in the piece:

One of my favorite projects (conceptually) that hasn’t been brought to life is kinetic art based on moire interference patterns. When sets of lines overlap, interesting visual patterns arise. When motion is added, the visual effect is extremely dramatic. In the spirit of works by David C. Roy, I’ve always wanted to make my own kinetic art based off of these patterns.

My conceptual workflow has been to create shapes in Inkscape, and then use Processing to animate the shapes to determine their visual appeal before I go and build them. From there, it’s easy to turn an Inkscape vector graphic into a .dxf file for rapid prototyping (water jetting for metal, router for wood, laser cutter for cardboard/plastic) and then we’re off to the races!

Below are some of the concept shapes I came up with, along with their respective animations.

This shape (called ‘simple’) is a pretty straightfoward shape with radial symmetry. The radial lines are curved to the give the shape a handedness so I can create two antisymmetric shapes to overlap with each other. The two overlayed shapes can be rotated in opposite directions to create a scintillating interference pattern (see center box). The rightmost box shows a pattern created from taking two copies of the same pattern and overlaying them rotating at different speeds. This causes each arm to appear to grow thicker and thinner in a seemingly periodic fashion. Click on either of the images to the right to see the animation; the original .svg file is available for editing (or for transferring to a .dxf file) by clicking on the image.

This next shape is an experiment with a shape that doesn’t exhibit radial symmetry. This pattern is more reminiscent of moire patterns induced by sets of parallel lines (there Wikipedia article linked above has some great examples of more traditional moire patterns). These shapes are achiral, so there’s no way to make them rotate antisymmetrically. However, just rotation in opposite directions gives an interesting visual effect of the shapes expanding and contracting with varying speeds. As before, the original image is on the left with the animation linked on the right.

This is my favorite shape so far, which came out of trying to make a round shape that was asymmetric but chiral. Antisymmetric copies of this shape rotating in opposite directions gives rise to a very complex visual pattern that has different feels to it at different points of the rotation. I’m a huge fan of both modes and the different ideas they evoke.

I’m looking forward to seeing how these turn out once they’ve been turned into real objects — I’ll be sure to post updates when that time comes

It’s a bit computationally intensive, so I should think of a better way to update the pixel information. It starts to slow down substantially at around 400 x 400 pixels, and I’d like to have it at 512 x 512 if possible. If you have any thoughts on a speedier algorithm, let me know!

I’ve been on a bit of a Processing spree lately. This week I continued my previous adventure in noise by exploring two-dimensional noise maps. The short version of what I did was turn every x,y point in the plane into a single perlin noise value, and then feed that value into another function that does something interesting with it.

The first thing I did was make that perlin noise value proportional to the intensity of the color in each pixel. That was pretty boring, so I made one color proportional to the noise, and the other color inversely proportional to the noise (so one color dominates for high values of noise, and another color dominates for low values of noise). That resulted in an intensity-based visualizer:

Intensity-Based Perlin Noise Visualizer

The next visualization I tried was to divide up the grid into 5×5 segments, and each segment was filled with a box proportional to the magnitude of the noise. Bigger noise, bigger box; smaller noise, smaller box. In addition, the shade of purple varies with the noise.

Size-Based Perlin Noise Visualizer

The last one was an attempt to create something a bit more visually complex and appealing than the previous attempts. I take each point in x,y space and make it the center of a circle. The value of the noise gets mapped between zero and two pi, and a line is drawn between the center of the circle and a point on its circumference that is at the aforementioned angle. This causes each line to rotate depending on the value of the perlin noise at each point. Sound a bit confusing? Perhaps an example will help:

Angle-Based Perlin Noise Visualizer

There are a couple of different directions that I’m planning on taking with this. The first is to animate it — I think that some of these visualizations would look quite pleasing if I animated them to shift in a subtle manner. The other potential direction is to use them as photo filters, as I did with the differential/proportional filters. I think the ‘blocking’ algorithm could be used to quite an effect on photographs, to distort the original image and create a new work from it. The rotational algorithm could also have such an effect. Hopefully I can draft up some new photo filters this weekend and test them out! There is definitely more to come.