Today, we had a meeting with Linda and the rest of the 2D group in 114 Sieg. We discussed the project a little more. I got the XML book from Sal and explained after the meeting the way in which I envisioned the data structure of the objects in C++. I think we should have an image types that has separate subfields for the structure, texture, and color data that we will get from the XML files generated by Yi's code. I am going to contact Yi, now that he has returned to Boston from vacationing, once I have a better idea as to the way I think his code works. I should really do this by Wednesday.

We also attended the little workshop for the Microsoft interns today. The presentations were amusing. I introduced myself to Raj Rao, after his presentation, and intend to email him to set up an appointment to meet within the next couple of days. His work seems mostly overlapped with my interests. In addition to him, though, it seems that the EE department here houses a lot of the areas that I would like to explore. For instance, I found out that most of the speech recognition and natural language processing occurs in EE. Also, the Human Computer Interaction people are housed in EE (In the HitLab, as previously mentioned in my notes). I should ask Linda how easy it is to work with the EE people while housed in CS.

To do:

1. Start writing code in C++ for the data structures.
2. Put up links to the references from the Fergus paper.
3. Figure out what HSI is exactly.
4. Contact Yi about the tags.
5. Start looking at the XML book from Sal to figure out the tagging of each image.
6. Make the actual layout for my webpage using cascading style sheets.

We had lunch with the group of DMP girls today, by 15th and 42nd for lunch. It was interesting to hear about what the other people were doing. I was interested in hearing Ruth, because she is just finishing up her PhD here at UW and is currently lecturing at UVA, I believe. I should ask Linda who she is exactly and what she does exactly.

I met with Jenny in the afternoon to go over what needed to happen for her work. She mentioned that Clifford was "organizing the database," and I got nervous about that idea, so we headed down to the lab and corrected the situation after talking a bit with Linda as well.

After meeting with Jenny for about 15 mins in my office, I had to do damage control with Linda. One of our office mates, who is a complete pain, complains at the smallest noise. And, so, he complained yet again. Linda showed me some places to meet with people outside my office. I hope those work out well.

I also talked with Linda about HSI and have a better understanding as to what that is about (hue, saturation, intensity).

Got a grant proposal from Viet today. Might want to check it out. I think it has something to do with Sal's work, but not totally sure. Might just want to ask Viet about it too. Thinking that I might be able to work on someone's grant. Should ask Linda about those sorts of opportunities too.

Literature search results yielded the following results (from the Fergus paper, entitled Object Class Recognition by Unsupervised Scale-invariant Learning):

Links of papers referenced by Fergus

To Do:

1. Read more XML and figure out Yi's stuff with it.
2. Read over the articles that are referenced, now that they have been compiled.
3. Figure out the EM algorithm with Arati, since both our projects use it.
4. Get my page online so that I can access all my links remotely, especially since I need to meet outside my office.

I met up with Clifford and Jenny again today to review the pictures that they have been getting. I assigned Jenny to the creatas page and Clifford to the freefoto page. They should have a number of pictures compiled for tomorrow's meeting.

Asked Jenny to link the list of articles from Fergus to our main 2D site. Hopefully she did that. I think Sal wants to talk about those references tomorrow, but I am not totally clear on that. Tried to get in touch with him late in the day, but didn't get through.

Talked with Sal about making the group dynamics more clear at tomorrow's meeting. Going to focus more on coding up the data structures in the next few days and worry less about Clifford's and Jenny's work. I am going to set office hours for the week, so that they can ask me questions. I am thinking maybe right after our meetings on Mondays and Wednesdays, both in the afternoon. I am also thinking about 4-6 hours per week should be sufficient. I will decide tonight and bring it up at tomorrow's meeting.

Today the 2D group had a meeting in which Jenny's and Clifford's roles were outlined more explicitly. They have been working on expanding the database, and they will continue doing so until they have enough photos for Sal to review. We will then evaluate our positions on each class that we would like to characterize.

To do:

1. Get prior notes off my personal computer and integrate them.
2. Figure out how to use the XML/C++ parser.
3. Understand Yi's notes on the organization of the tags in the different image files.
4. Explicitly map out the design of the data structure based on the understanding of those tags.

Yay! Day off for the fourth of July!!!:)