General

FACULTY

SCOPE

All G&V students are expected to know the basic graphic principles and techniques (lightfield, ray-tracing and z-buffer scan-conversion, view control, scene graph computation, rendering pipeline operations) and the basic 2D graphics and 3D modeling terminology (topology, geometry), representation schemes (CSG, BSP, Voxels, octrees, triangle mesh, triangle strips) and techniques (key-frame animation, intersection calculations, convex hulls, marching cubes). They should also be aware of the major issues (geometric complexity, realtime performance, image accuracy) and research opportunities in the overall GVC area.

SUGGESTED READINGS

In addition to these references, students should be familiar with the broad topics covered by papers in the recent proceedings of the ACM Siggraph and of the ACM Symposium on Interactive 3D Graphics.

SUGGESTED COURSES

SAMPLE QUESTIONS

  1. GVC SUB-FIELDS: Define the following terms, explain how the differences between these areas and how they relate to each other: 3D graphic, photo-realistic rendering, image-based rendering, animation, visualization, vision, virtual reality, augmented reality, computational geometry, graphic information systems, computer-aided design. For each, describe the fundamental technologies that play an important role, the principal applications, and major research issues.
  2. BENEFITS OF INTERACTIVE 3D: Define the term "Interactive 3D graphics", explain its objectives and its benefits, discuss its technological and sociological challenges. Explain precisely the benefits and drawbacks between the following three desktop metaphors: (1) the current Microsoft Windows; (2) a system where each window is presented as a static photograph of a 3D office where other folders appear as desk drawers or books on shelves; and (3) an interactive 3D navigation system showing a large virtual office with open doors to other similar offices and lots of books and devices on shelves, which, when approached, open files and start applications. What would you do to make the third metaphor really effective? What would be the major research issues? How would you market the result?
  3. ELECTRONIC COMMERCE SOLUTION: Suppose you were to design a real-estate system, where buyers could use on on-line system to interactively walk through photo-realistic virtual 3D models of houses for sale from their home. You must minimize the cost of data acquisition and the waiting time for the buyer, and maximize the graphic performance, the ease of use, the appeal of the imagery, and the feeling of freedom to tour the house at will. Describe in details such a system and explain how you would go about building it, including the off-the-shelf software, hardware, and peripherals.
  4. OPEN RESEARCH PROBLEMS: List the 5 most important open problems in the Geometric and Visual Computing area that you expect to be funded and researched actively over the next 10 years. Discuss the motivation and the current state of the art. Explain the research issues and say what makes them hard. How would you attack them if you had to?
  5. COMPARE REPRESENTATIONS OF SOLIDS: Describe the following five representation schemes for solids: Boundary representation, CSG, BSP, Voxel, Octree. For each, describe in broad terms a typical data-structure and give the details of an efficient algorithm for classifying a point against a solid. Compare these five representations in terms of storage, performance of point-classification, and convenience for interactive design.
  6. PERFORMANCE ACCELERATION: Discuss the stages and factors that limit the performance of current generation hardware-assisted rendering pipeline architectures. Present all the general purpose graphic acceleration techniques that you know. For each, explain in broad terms the principles, fundamental algorithms, data-structures, expected benefits, and limitations.
  7. IMAGE-BASED RENDERING: Describe the Luminograph and Light Field Rendering methods presented in SIGGRAPH 1996. Compare these methods and discuss the relative advantages of each. Which method is more general? Which is more computationally complex? What are their limitations.
  8. GROUP ANIMATION: Explain in details how you would program and animate the behavior of a flock of birds flying through a city between tall buildings. The birds must stay together while avoiding obstacles, yet each bird controls its own trajectory and does not get too close to its neighbors. Suggest efficient data structures for detecting collisions and for maintaining proximity information between birds as their nearest neighbors change.
  9. MORPHING: Consider two objects A and B, each represented by a triangulated boundary. Describe in details four techniques for animating a morphing operation between them. What does it mean for the morph to be smooth? What does it main to be minimal? For each one of the techniques you described, discuss their smoothness, minimality, and restrictions.
  10. ISOSURFACES: Describe in detail a technique for finding isosurfaces in a volume of scalar data. What is the computational complexity (e.g., in terms of the number of data elements) for the steps in this technique? Suppose interactivity was a main concern. Describe in detail hierarchical methods that might be applied to create isosurfaces in real-time or near real-time. If the data is time-dependent, how would this affect the method? Within the framework of your technique, how would you manage the presentation of greater detail? Describe both automatic and user-controlled methods.
  11. VISUALIZATION TECHNIQUES: The visualization of discrete volumes of data can be accomplished using three broad families of approaches: direct volume rendering, projection (or splatting), and 3D texture. Describe each type of approach, and highlight the differences between them.
  12. IMMERSIVE 3D INTERFACES: Present an argument for or against the following statement: "Interaction in virtual environments should be as close as possible to interaction in the physical world." Do you think that VEs will eventually have a "standard" interface, analogous to the desktop GUIs on today's PCs? Why or why not? List some advantages and disadvantages for each of the following methods of interaction evaluation: usability study, single-variable experimentation, testbed evaluation.