Ideas For Final Project

Don't forget to hit reload/refresh.

Name: Krishnan Sriram
E-mail: sriram@mit.edu
Idea: Myself and Ira Cooper have come together with an idea about simulating a billiards game. As we see it, it involves lot of physics, light effects and an enormous amount of computation!! We are looking for the third partner for our team.

Name: Valentin I. Spitkovsky
E-mail: val@mit.edu
Idea: I work in the Computational Biology group at the LCS. We have some algorithms that find related sequences (strings). Of course, there is a ton of data -- the databases of sequences often have millions of entries in them. It would be very useful to have a good visualization algorithm that would allow one to actually look at the related sequences, highlight the matches, zoom in on areas of similarity efficiently, etc. As of now, the algorithms simply dump huge text listings. If you are interested in biology and graphics, this should be a really cool project. It would certainly be very useful for biologists and the people in my group; there are many ways to be creative in the design to make the visualization intuitive, easy to use, and highly informative. There's no shortage of data to visualize. It will be interesting and challenging to find a fast and elegant way to deliver the important information in an understandable way.

Name: Joshua Glazer
E-mail: vbunny@mit.edu
Idea: I want to make an animation involving live video and 3d characters. I'm thinking we'll film a short (3 minute) kung fu fight (between one person and empty space) in DV and keep track of lighting and camera work. Then, we'll edit the video and add a 3d rendered character with which the live character interacts. We must match lighting and camera angles and take special care rendering areas where the two characters interact. I have access to necessary equipment.

Name: Frank Dabek
E-mail: frank@mit.edu
Idea: Weathered Surfaces - Rendered scenes often appear unrealistic to me because they are so clean. I propose to simulate the effects of weather on objects -- dust, grime, discoloration, scratches etc.

Name: Ron Dror
E-mail: rdror@mit.edu
Idea: My basic goal is to produce very realistic images of simple surfaces. In the real world, light strikes objects from every direction. If you look at a piece of aluminum, you see a distorted, blurred reflection of everything around it. Computer graphics renderings, on the other hand, tend to use very simplified models of lighting (and also very arbitary models of reflection, like this ray tracing assignment's Phong model that I hope to finish implementing some time this morning). I want to use a camera with a fisheye lens to capture the lighting incident at one location in the real world. This will then serve as the illumination for a surface, whose reflectance will be represented by a fairly accurate physical model. The result should be a surface that looks real. One nice thing about this project is that getting the basics to work should be fairly quick. I have already thought about this a lot, and I know how to do the basic computations for simple surfaces. However, there's no shortage of possibilities for improving it -- we could try to render in real time, add new surfaces and effects, work with more complex objects, and so on. FUN FUN FUN FUN FUN!!!

Name: Prof. T.
E-mail: teller@lcs.mit.edu
Idea: Prof. David Mindell has discovered some deep-sea shipwrecks in the black sea. He has some challenging data, including photomosaics, side-scan sonar, and sub-surface sonar. Visualizing all of this data together -- for example, by texture-mapping some kind of volumetric representation, with all three data streams aligned in 3D -- would be a good project, and could lead to future UROPs, etc.
Contact Bill Wallis if interested.

Name: Flavia Sparacino
E-mail: flavia@media.mit.edu
Idea: A CHINESE SHADOW THEATER Mark Huang and I are joining efforts on a chinese shadow theater project. The purpose is to use hand input through computer vision and create the corresponding 3D character which morphs out of the shadow. We are looking for a 3rd team member who has possibly an artistic inclination and is more willing to work on the 3D modeling (alias or maya) aspect of the project.

Name: Shishir Mehrotra
E-mail: shishir@mit.edu
Idea: I am interested in doing something related to Computer Vision. One idea is to create some kind of interactive environment where movements in the real world are mirrored by movements in the object space. I am open to other ideas, let me know....

Name: Alice Yang
E-mail: ayyang@mit.edu
Idea: MTV Lei and I are thinking of creating animated Music Videos for our favorite songs. We'll be creating characters that can "dance" to the rhythm. This'll be a fun project, and we're looking for a third group member. we're very open to any new ideas too.

Name: Jaroslav Blagojevic
E-mail: jar@mit.edu
Idea: Organic forms are today usually modeled as any other--by specifying vertices, surfaces and normals. We want to make a system in which we specify the "DNA" of the object, and then watch it grow from a simple seed (examples: a face, a plant, a creature, and assignment3 jar_glock.iv). The idea is similar to the idea that the body of every living organism grows from simple stem cells. An OpenGL based experimentation/visualization system written in C++ is already available.

Name: Henry Wong
E-mail: henryw@mit.edu
Idea: Chris Avrich (cda@mit.edu) and I are planning on modeling a large network of boxes connected by springs. The project would involve doing a fair bit of physics modeling as well as various lighting effects etc. The end result would be a program that lets you construct an arbitrary network of boxes connected by springs of various stiffnesses. You could pull one of the boxes out of position and the program would simulate how the entire network of springs would oscillate.

Name: Hai Ning
E-mail: ninghai@mit.edu
Idea: How about a networked modeler. Mutilple user can interactively model same scene. They can also have chat, shared white board. At some point, they can save their work to a database, and retrieve later. What's better, the database will even record their design process and later on, it could be replayed for designer to better understand each other's idea. This could be done in Java applet with VRML (very similar to Open Inventor).

Name: Peggy Kuo
E-mail: peggykuo@mit.edu
Idea: Hi, Charles and I teamed up with an idea at the following URL: http://www.media.mit.edu/%7Epeggykuo/6.837/final/proposal.html; it's about to build a 3D character which can have certain motion according to user's input. I know it's a little late, but we are still looking for one or two partners. If you are interested in Game programming, character modeling, our idea, or are still looking for idea matching persons. Please email peggykuo@mit.edu.Thanks very much! We need partners!!