Figure 1: 3D Photography examples: Beauvais Cathedral model; Mt. Polizzo site model; Le Penseur sculpture model; Fort Jay, Governors Island, NY.

Class schedule, syllabus

Project 1, Due Feb. 25

Project 2 (Proposal due Mar. 11)

Project 2 Presentation and Report Guidelines

Project Teams

The ability to acquire, view, animate, and analyze 3D objects in a computer is becoming increasingly important. When we see 3D objects in a computer, we need to ask the question "how did they get in there?". There are many different methods and approaches to answering this question. The methods differ depending on size (how do you acquire and model a molecular structure, an anatomical part of the body, a person, a machine, a building...?) and function of the model (model's that need to be viewed only do not have to be geometrically or topologically correct; models that are used for physical analysis or simulation may need to be). This course will explore 3D Photography - the process of automatically creating 3D texture mapped models of objects -in detail. Applications of this technology include robotics, medicine, graphics, Virtual Reality, entertainment, digital movies etc. We will discuss 3D data acquisition devices, 3D modeling systems, and methods to acquire, create, augment, manipulate render, animate and physically build such models.

The course will have two complementary components. Weekly lectures will introduce the topics above, and will consist of reading a set of research papers on each topic, along with a class discussion. A class schedule and course syllabus is available online. The second component consists of group projects centered around using 3D photography to model objects (buildings, rooms, people, objects, etc.). Students will have hands-on use of a variety of scanning equipment in the Columbia Robotics Lab,including 3D pritners for modeling.

PREREQUISITE: Experience with at least 1 of the following topics: Computer Graphics, Computer Vision, Robotics or Computer Aided Design or permission of the instructor. Students co-enrolled in a vision or graphics course during the semester may be able to take the course. Students need a proficiency in at least one of C/C++/JAVA languages. Any student who is unsure of his/her background should contact Professor Allen.

INSTRUCTOR: Prof. Peter Allen. Office: 619 CEPSR. Phone: 212-939-7093, allen@cs.columbia.edu. Office Hour Mon. 6-7.

TA: tba, .

TEXT: Papers and handouts available online here.

COURSE FORMAT:

The class will meet once a week with lectures by Professor Allen. The grade will be based upon the following: 50% for class participation and 50% for group projects results.

TOPICS TO BE COVERED:

• Overview of 3-D Modeling
• Acquisition devices: Range scanners, stereovision, robotic digitizers etc.
• Registration Methods
• Representations for 3-D objects
• Creating 3-D models from multiple Range Views: Surface Models
• Creating 3-D models from multiple Range Views: Volumetric Methods
• Passive Vision Methods: Stereo, Shadows
• Space / Volume Carving
• Texture Mapping 3D Models
• Viewpoint Planning
• Applications and 3D Scanning Projects
  • Archeaology
  • Urban Planning
  • Architecture
  • Medical
  • Digital Movies