Jump to a Certain Week:

My Research Experience

A Weekly Journal

Week 0

The first person I met on campus turned out to be the other DMP student I worked with over the summer, Noor Martin. Most of the first week was spent getting accounts set up, keys for the labs, and U-Cards. Noor and I quickly familiarized ourselves with the large campus by trekking across campus *multiple* times to set up our university accounts and meet our mentor, Victoria Interrante. The process of getting things done at a large school makes you appreciate coming from a small univerity where you don't have to walk great distances and you can talk to the person in charge. On the other hand, we didn't have to worry about exercising.

The graphics group just moved from the EE/CS building to the Digital Technology Center (DTC) on the fourth and fifth floors of the newly renovated Walter Library. So our work for the first week included reading papers while the network was still being setup.

On Tuesday we had a meeting for everyone on the visualization project. Saturday evening there was an open house at Walter Library. Noor and I presented our project and met other graduate students from other areas in the computer science department.

Week 1

I started looking at the previous version of the code this week. Based on the meeting last Tuesday it looks like the areospace people are looking for improvements in the code and the addition of a GUI. After familiarizing myself with MUI, a OpenGL toolkit for C, I requested requirements of the interface at this week's project meeting.
Back to Top

Week 2

By the middle of the week I had created options to change the color of the vorticity and velocity patches. On Wednesday I added more buttons to distribute the functionality of the colorization options. Then on Thursday the real fun began. Some of the button presses resulted in a segmentation fault. The seg fault did not occur predictably and it could not be repeated regularly based on the same set of actions. By Friday I was out of ideas. Thankfully, Songho Kim, another grad student who had given us a briefing on MUI suggested using a debugger, ddd. With ddd we were able to pinpoint the area where the seg fault occured in the MUI code.
Back to Top

Week 3

Week number two with the seg fault. The problem has us all stumped. I determined the problem is not with the MUI code since there is no documentation on the web. In addition, I took out all the MUI code out of the program, adding it back in slowly only to get a seg fault at some random point. Next, I tried reversing this process by adding code to the interface code. Unfortunately, the size of the program and dependency of functions on each other did not make this a very effective method for debugging. On Friday I went over the code with Vicki. At this point she told me to continue working on the gui as long as the seg fault does not become a frequent occurence.
Back to Top

Week 4

I emailed some of the professors in the department this week. Other programs at the university have seminars to highlight certain areas in their department so I decided I should do the same to find out more about the computer science department. Coming from a small school I have not looked into some of the areas available at a larger university. So far I have set up a meeting with Dr. Pen-Chung Yew , the head of the Electrical Engineering department.

I've added code to allow the user to change the way the region of interest is desplayed. There are breakpoints to change the threshold when displaying the vorticity patches. I also created two graphs for a graphical view of the vorticity colors displayed on the visualization. The graphs are especially cool because the user can click on the breakpoint and drag it to the setting they desire. It allows the interface to be more interactive and easy to understand since the color of the breakpoint matches the color shown on the visualization.
Back to Top

Week 5

In addition to working on the visualization project I started working with Haleh Hagh-Shenas on her dissertation topic texture morphing. I am looking forward to working with Haleh since it is an unsolved problem and it will be a new challenge.

I found myself at the paper stage once again. I started working with Matlab at the same time to see what tools it has for manipulating images. I have also met with Haleh regularly to discuss her findings up to this point in time and to ask her questions I had over the readings.
Back to Top

Week 6

After learning how to use Matlab I tried several ideas that did not prove to be useful. First off, it would be nice to determine the edges of the texture, but if you filter images, you also filter out the edges. Next, I tried increasing the contrast of the texture. The problem there is when the texture's color interferes with edge detection. Even with these dead ends, I am encouraged. Maybe I have not found a solution, but I *know* that something doesn't work because I tried it.

With this new understanding for the level of difficulty involved in texture morphing I decided to simplify the subject by taking a simple case of binary objects. I started coding a program to read in the texture, identify objects of similar color, and getting the initial texture to linearly move towards the destination texture. Next week I would like to work on an algorithm for altering the intial texture objects angle in space while it moves towards its destination, allowing the morphing to look more realistic.
Back to Top

Week 7

Both Vicki and Haleh are at SIGGRAPH this week so I started adding Noor's changes of the visualization code to my version. I presented my interface to the aerospace people for the first time, and they encouraged me to add the vorticity functionality to the other measurements. Therefore, I am continuing to create more buttons on a daily basis. So far my record is still 26 buttons in one day.
Back to Top

Week 8

I continued to develop the functionality of the program while trying to make as few mistakes as possible. I discovered that the MUI radio buttons and textboxes interfere with each other even when they are invisible. Secondly, the different types of buttons react differently to the same situation. For example, the first radio button created can be seen, but another button at the same location will have interference from the first button. Meanwhile, textboxes will show up no matter what, but only the last textbox initialized will acknowledge input. I'll have to redo the layout of the interface after I get the visualizations to appear correctly.
Back to Top

Week 9

I finished the last addition to the code on Tuesday and moved the buttons around on Wednesday so all the options can be seen. Of course, no project would be complete without documentation, so I wrapped up the project by writing a bit for future programmers and users who will work with the program.
Back to Top

Main -- People -- Research -- Life Outside the Lab