CS 348C - Topics in Computer Graphics

Sensing for graphics


The paper

Topic due Thursday, January 16
Paper due Thursday, January 30

Overview

The purpose of writing a paper is to familiarize yourself deeply with a research topic of your own choosing that is relevant to the course. We offer (below) a suggested strategy for choosing a topic. You are expected to research the topic and perform a synthesis of your readings. In particular, you should address the following questions:

Requirements

First, choose a topic and email us by January 16 a short description (one paragraph is fine) of your topic. Feel free to stop by during office hours and discuss topics with us. For some of these topics, we can offer you a list of published papers. For others, you will need to perform your own literature search.

The paper itself is due on January 30. The required submission format is an .html file. We prefer if you put the entire paper in a single .html file, rather than building a tree of .html files. We're not looking for theses here, just papers. 10 pages is about the right length. Feel free to include figures and images (of your own or from papers you have read, but don't go overboard). For inspiration, look at the papers submitted for 348C last quarter. The paper by Tor Norbye is particularly nice.

On the due date, email us the URL of your paper. We'll construct a table of contents and place the papers online for all students to read. On Tuesday, February 4, we will have a general discussion of these papers, and then we will set an agenda for student-led presentations and discussions of these topics, which will occupy us for the subsequent 7 class meetings.

We'd like you to work alone on the papers. However, we'll encourage you to team up for the presentations, especially if your paper topics were similar. You are also encouraged to team up for the final projects. We are flexible on the size of these teams, although we envision teams of 2-3 people; let's see how things evolve.

Choosing a paper topic

We suggest the following strategy. First, look over the list of suggested projects. If you have one in mind that isn't on this list, come talk to us. Make a short list for yourself of projects that might interest you later in the course. Now identify the intellectual core behind each project. This becomes your paper topic.

Identifying the intellectual core behind a project is not as hard as it sounds. In most cases, the core is obvious. For example, if you're interested in synthesizing tilable displacement map textures from range data, then a good paper topic would be to survey the literature on texture analysis / synthesis. If you're interested in image-based modeling of architectural environments, it would be reasonable to survey the state-of-the-art in shape-from-stereo.

To further help you, here is a list paper topics that each reflect the intellectual core behind one or a cluster of the suggested projects:


levoy@cs.stanford.edu
© 1997 Marc Levoy and Brian Curless
Last update: Sunday, 19-Feb-2006 23:11:14 PST