Shawn Lankton Online

The super short answer: It’s what I study at Georgia Tech.
The longer answer: below…

As part of a program I’m involved in I needed to describe my research succinctly and from a very high level to forty of my colleagues. Many of the people who I was addressing did not have a technical background. I drafted the following short essay to describe what Computer Vision is, and what I do with it as of right now:

Computer Vision: A Summary

Sight, in my opinion, is the most incredible of our five senses. I conduct research in a field known as “Computer Vision.” This area focuses on the higher-level parts of a fascinating overall problem: Teach computers to see like people.

Teaching computers to see can be considered in three layers. The first is image acquisition. Here, I use the term image to mean data that a computer can interpret. This includes pictures like those from a camera, three-dimensional brain scans, and much more. These images are fantastic tools for people, and people alone. Computers cannot understand them without something more.

The next phase is image processing. This step makes the acquired images ‘nicer.’ This includes signal processing problems like removing noise, enhancing colors, and making important features more apparent. A processed image, though, still holds no meaning for the computer… only the humans that look at them.

The third layer is where my research is focused. The computer vision layer takes these processed images and assigns meaning to them. Three key activities here are segmentation, registration, and tracking. Segmentation involves finding the boundary of an object or objects of interest in a scene; registration is the process of lining up two images; and tracking is the process of determining the position of an object over time in a video sequence.
Read the rest of this entry »

I want to share the lessons I learned recently when writing a paper. These were some revelations that helped me get over the procrastination hump and really set me on the writing fast-track.

1. Start with a thesis
2. Do all the experiments next
3. Recursive outlining
4. Finish it up

Read on for an explanation of each of these steps:

Start with a thesis

Before you do anything else, write a short, to the point thesis topic. It should be one or two sentences and the entire paper should be written to prove that thesis. I stuck mine to the wall above my desk to help keep me focused. They tell you to do this in grade school, and when you’re all grown up, the same rule applies. This will forever-on be the first thing I do when I sit down to write a paper.

Do all the experiments next

All of your experiments should be designed to support your thesis, and completed before you do anything else (so that you know your thesis is right). If your thesis turns out not to be right, then you have to go back to the beginning and pick a new thesis!

When I say experiments, here I mean the work that you’re writing about. In my case ‘experiments’ are figures showing the results of my computer vision algorithms. For someone else it may be analysis of a client’s financial data or a computer simulations of particle movement through a turbulent fluid field.

In any case, doing these with the thesis in mind, and before writing ensures that the experiments are relevant, and help to prove the thesis.

Recursive Outlining

In recursive outlining I start with an outline of sections: Introduction, Background, Novel Algorithm, Results, Conclusion. This takes no time because its the same for almost every paper. Next, go though and add sub-sections, then sub-sub sections.

At this point, most typical ‘outlining’ stops. However, keep going. Add sentence ideas to each sub-sub section, and start making rough guesses about where graphics will go. Add key words to those sentences ideas. All the time, keep the thesis in mind.

Now your paper has all of the ideas you want to present in a very rough form. Inevitably some sections will be sketchier than others, but now the process of refining the text can become very compartmentalized without sacrificing overall idea of the paper.

Finishing it up

What the process until now has done is to take apart the paper into these little tasks. The final steps are to turn your sketch for each sub-sub section into an concise, eloquent-sounding, finished piece of writing. Once you have done this for all the sections, give it a final read-over and its done.

Keep in mind that some of these steps may take days, but by following this process you always know what your next action is, and you don’t have to waste a lot of time doing pointless things or re-doing things that were directed towards the wrong point.

It’s true that I have precious little paper-writing experience, but these revelations have helped my writing considerably. I plan to post updates to this as my experience grows and my method evolves. If you have any good tips along these lines, please share them in the comments!

In the past few years, I have discovered a method for taking notes that makes the notes easy to use, and turns them from a pointless exercise in dictation to a useful reference! I’ve going to “school” for some time now (about 18 years). Of course, for much of that time I was diligently copying down the words spilling from the prof’s mouth so I could study it later. I started experimenting with different note-taking styles a few years ago, and have finally settled on something that works. With a little planning and a few extra minutes of organization you can turn a course worth of information into a great reference. I use notes from classes daily now! Read the rest of this entry »

The sands of time can be slow to fall through the hourglass, but today some key grains made the drop into illustrious lower pile! On this day, I found in those grains the time to post a few more old projects into this new incarnation of the website. Check these out if you haven’t already seen them:

Keyboard Snooper
DTMF Generator
Java Minesweeper

I also added the download-able content back into the other projects. Some of the other projects (namely Visual Tracking) are due for a re-write any day now.

Some days you just jump up and down and then do a little boogie-woogie right in the office. Today was one of those days. The day began with a slow start and some celebratory end-of-semester lunches. When I got back to the office and checked my email…

I found out that I got a TI:GER fellowship that will let me work on cool entrepreneurial projects with MBAs and law students from Emory. (*AND* give me a $7,000 per year raise!)

Ok, I took that in stride, and played it cool. Then I got to work on that off-frame tracker I’ve been talking about… I added a non-hacked version of on-line motion model learning, and got some stellar results! That did it. I started dancing. Too bad no one was here to see : )

Its days like this when I really, really, *really* love my job.

I’ve been working lots this past week… You could say I’m working too much because I’m choosing to work on research rather than do homework and write papers for my classes… oops! Anyway, the object of my obsession is some cool new visual tracking algorithms. I’ve been terribly remiss in putting up some details on my projects page, but in the mean time I’ll taunt my internet audience with some results I presented to my group at a meeting this Thursday: enjoy!

These are some people walking around outside my office.

Don’t know much about this video, but I’ll bet those are cars.

They were all tracked with a technique called “Kernel Tracking.” I can get these results pretty fast (real-time). Right now I’m working on some even cooler stuff with “Particle Filters.” More coming soon.

I guess this happens to anyone in research, but it happened to me today…

The build-up: You’ve been reading the literature, you’ve come up with nick-names for the relevant authors in the field, you feel ‘with it’ enough to make snap judgments about the papers that are clearly no good. Then the proverbial light bulb blinks on! “What a good idea I’ve just had,” you think. You even go so far as to tell your friends to demonstrate how clever you are!

The discovery: Now that you’re on to something hot, you start reading even more literature to get help with the techniques and see what kinds of results you have to beat to really wow the community. Suddenly… low and behold… damnit! Someone has already had, implemented, and published your idea! Alas… at least the idea worked, right? Now I just get to implement it, but no publishing allowed… oh well.

I just finished making the poster that I will present in San Diego next week (take a look). I have a couple of thoughts regarding this experience:

1) Vector graphics are soooo cool. Always use them if you can… always. (note: I talk about how to get vector graphics in .eps form out of Matlab quickly and easily here: Matlab TeXniques)For those of you who don’t know why vector graphics are cool, let me enlighten you! When you resize a normal image (gif, png, jpg) you will run into problems (click image above). If you make it smaller, you throw away data, and if you make it bigger, you have to make up data! Both of these are bad. With vector images, though, you can scale them up and down all you like and they look as crisp & clean as the day they were created! Wonderful.

2) If you want to put LaTeX symbols in your power point file, forget TeXPoint… All the cool kids are using TeX4PPT these days. Its *free,* much easier to use, and creates (you guessed it) vector images of your LaTeX stuff instead of bitmaps. One potential hang-up is that it only runs on Office 2003+. Sorry all you guys still on XP.

Some slight problems that I had: I think it was due to not setting paths right, but I had to put my .ppt file in the miKTeX root director for it to compile the TeX. Also, you can’t right-click your text boxes to “TeXify” if the text has the squiggly red underline (just find an un-underlined place).After that rousing adventure in technical advice… if you’d like to see my poster without flying all the way to San Diego, click on the image above to open it in .pdf!

These bad boys model your “target” (maybe a football player) by a histograms in a little window around the target. Then, they try to find where that window should be centered in the next frame to have a similar histogram (i.e. be lined up with the moving target).

Its a very simple idea and it seems to be working well. And, guess what else? This paper original came from … SCR (yup, the very same building… in Princeton, NJ). Here’s the link to the paper if you’re interested: Kernel-based Object Tracking. Enjoy. Hopefully I’ll have some cool results of my own, instead of lifting them from someone else!

I have just received notification that my third academic paper, “Hybrid geodesic region-based curve evolutions for image segmentation,” has been accepted to the SPIE Conference on Medical Imaging. That means that I will be traveling to San Diego in February of 2007 to present it.

I will be listed as first author on this paper which is a big distinction. It means that the idea and implementation were primarily my work. Furthermore, grad students are more or less judged on a basis of how much they publish, so the more the better. This will be the first publication from my current group at Georgia Tech! Maybe I should add a “publications” section to my webpage?