CSE 581: Parallel Computation in Image Processing

Welcome to the 581 Home Page!

This is the World Wide Web ("the Web" for short) hypermedia document for CSE 581 and contains information about the class. Keep in mind that this document is not static, and that new information will be added from time to time.

Schedule Information

On October 31, November 7, and 9, class will begin a half-hour earlier than normal. I.e., we will start at 8:30 on these days. On November 1, we will meet at 9:00 and have a guest speaker, Prof. Ze-Nian Li of Simon Fraser University in Burnaby, B.C., Canada.
Tuesday, November 14: no class.
November 16 at 9:00. Dr. Bharath Modayur will be our guest speaker, and the title of his presentation is, "Efficient parallel object recognition on SIMD and MIMD machines".
Tuesday, November 21. Class will begin at 8:30 AM. Topics: Completion of discussion of pyramid algorithms -- scale-invariant operators, top-down algorithms, and segmentation through hierarchical relaxation (using the ISODATA approach of Burt, Hong and Rosenfeld). Introduction to embedding and virtual processing. Overview of neural network architectures. algorithms,
Tuesday, November 28. Class will begin at 8:30 AM. Topics: Completion of overview of neural network architectures. Embedding of neural networks in meshes and pyramids. Brief treatment of iconic/symbolic computation.
Thursday, November 30. Class will begin at 8:30 AM. Topic: Parallel image analysis for digital libraries.
Here is the demo schedule.

Finding term project topics

During the week of October 12-18 students should be actively exploring one or more topics for the term project. Written descriptions of topics should be handed in on Tuesday, October 24. A template for the writeups is available here.

Resources

PVM (Parallel Virtual Machine) is a software layear that permits a user program to run on a virual machine made up of a heterogeneous collection of one or more workstations. This is a convenient way to implement and study distributed algorithms.
Intel SSD Technical Publications include documentation for the Intel Paragon parallel computer system.
The ZPL language is a good language in which to implement 2-D array-oriented algorithms on the Intel Paragon.
Various vendors of supercomputers and parallel machines.
Some info on the MasPar, from the National Supercomputer Center in Sweden.

Some online information for the MasPar MP-2 is at the University of Tennessee. This resource was found by Neal Friedman. He reports that "they've also got some Paragon documents."


Here are some errors in and corrections to the course notes.
Copyright Notice: The material in this course web is subject to copyright. While it may be viewed by the public, it should not be installed at any web site other than the one at the University of Washington.

Term projects are an important part of the course. These should be started during the week of October 16.
The review session for the final is schedule for Friday, December 1, 4:30-6:00 in 422 Sieg Hall.
The final exam is scheduled for Wednesday, December 13 from 10:30 to 12:20 in our normal class meeting room. The exam is closed-book.
Term projects are due at or before 4:30 PM, Thursday, December 14.


(Last Update: 10 October 1995)