EE292B Course Information
EE292B -- Electronic Documents: Paper to Digital
Course Information
Spring Quarter '04-'05
Meeting place and time:
Hewlett 200
, Tues and Thurs 4:15-5:30
Lab Session:
ISE Laboratory, Thurs 6:00-8:00
Goals of the course:
To provide a basic understanding of the concepts and principles
that underlie techniques of acquiring, storing,
and accessing paper documents in digital form, for use in personal and
corporate collections and digital libraries.
Prerequisites:
Basic probability, linear algebra, and programming skills;
EE103 or EE168 helpful but not required.
Instructor:
Ashok C. Popat
Palo Alto Research Center (PARC)
3333 Coyote Hill Road
Palo Alto, CA 94304
Phone: 650-812-4899
popat@parc.com
office hours: Tuesdays, end of class until 7:45pm, ISE Lab
With help from:
Dan S. Bloomberg
Google
Phone: 650-623-5004
dbloomberg@google.com
office hours: Thursdays, end of class until 7:45pm, ISE Lab
Course Secretary:
Kelly Yilmaz
yilmaz@stanford.edu
Packard Building, Rm.259
650-723-4539
Text:
Managing Gigabytes, 2nd Ed. by Witten, Moffat, Bell.
ISBN: 1-55860-570-3. Estimated price: $70.00
Supplemental Readings:
Additional readings will be handed out in class. Readings for which
the copyright statement or permission letter allows it
will also be made available electronically to
registered students, on or shortly after the handout date.
Handouts:
Many of the handouts will be made available electronically to
registered students, on or shortly after the handout date.
Copies of handouts will also be available for pick-up after the handout date
in the course drawer in the cabinets on the second floor of Packard.
Grading:
Problem sets: 60%
Project: 40%
Students are expected not to collaborate substantively
on problem sets. For the
final project, students may collaborate in teams of up to three.
One homework may be handed in late without penalty.
Details about what's expected in a project, along with some suggestions
for project topics, are available
here.
3/29 Tues: Course overview; review of basic signal/image processing;
digitizing the world's books (slides)
Material covered: MG Chapter 1, Chapter 10
Handouts: Notes on Document Image Analysis,
Burrows-Wheeler tech report
3/31 Thurs: Data compression basics, Huffman and arithmetic coding (slides)
Material covered: MG 2.1-2.4; omit material marked with gray bar
Handouts: Problem Set 1, Haralick et al. morphology paper
4/05 Tues: Burrows-Wheeler, Ziv-Lempel; Document image processing: deskewing, binarization, connected-components analysis (slides)
Material covered: MG 2.5-2.6 (omit DMC, PPM, and material marked with gray bar), 7.3, 8.1, Notes on Document Image Analysis
Handouts: Notes on Inverse Halftoning, Project Guidelines/Suggestions
4/07 Thurs: Document image processing (cont.); Morphological Image
Processing (slides)
Material covered: Notes on Document Image Analysis, Haralick et al. morphology paper (except grayscale)
Handouts: Notes on Hausdorff Distance, Basic Definitions in Morphology
4/12 Tues: Morphology (cont.); Layout analysis / segmentation (slides)
Material covered: Haralick et al. morphology paper (except grayscale),
MG 8.2
Handouts: Problem Set 2, Jung and Seiler paper
Due in class: Problem Set 1
4/14 Thurs: CCITT Fax, Context-based, and JBIG Image Compression (slides)
Material covered: MG 6.1-6.4
Handouts: Solutions to Problem Set 1, Kopec and Chou DID paper
4/19 Tues: JBIG2 and Mixed Raster Content (slides)
Material covered: MG Chapter 7 except 7.3, Notes on Inverse Halftoning
Handouts: Hobby and Ho paper, Bern and Goldberg paper, Greene et al. preprint
4/21 Thurs: Brief intro to pattern recognition, Markov-model-based Text Recognition (Document Image Decoding) (slides)
Material covered: Kopec and Chou DID paper (except scheduling)
Handouts: Problem Set 3, Notes on Statistical Lexical Language Modeling for DID
Due in class: Problem Set 2
4/26 Tues: Super-resolution of document images (Guest Lecturer: Marshall Bern) (slides)
Material covered: Bern and Goldberg paper
Handouts: Solutions to Problem Set 2, Notes on EM Algorithm
4/28 Thurs: Statistical Lexical Language Modeling for DID (slides)
Material covered: Notes on Statistical Lexical Language Modeling for DID, MG 2.5 (PPM)
Handouts: Jain et al. paper on clustering
Due in class: Project proposals
5/03 Tues: Coding Data on Paper to Improve OCR (Guest Lecturer: Dan Greene) (slides)
Material covered: Greene et al. preprint
Handouts: Hofmann SIGIR99 paper
5/05 Thurs: Basic Information Retrieval, Vector space models, Latent Semantic Analysis (slides)
Material covered: MG 3.1-3.3, 3.7, 4.4, 4.5; omit material marked with gray bar
Handouts: Solutions to Problem Set 3,Problem Set 4
Due in class: Problem Set 3
5/10 Tues: Perceptual Document Analysis (Guest Lecturer: Eric Saund) (slides)
Material covered: TBA
Handouts: Chen-Bloomberg DimSum paper, Kupiec et al. SIGIR95 paper
5/12 Thurs: Mixture Models and Expectation-Maximization, Clustering (slides)
Material covered: Notes on EM Algorithm, Jain et al paper
Handouts: Reading on e-books (details TBA)
5/17 Tues: Probabilistic Latent Semantic Analysis (Guest Lecturer: Thorsten Brants) (slides)
Material covered: Hofmann SIGIR99 paper
Handouts: Reading on distributed digital libraries (details TBA)
Due in class: Problem Set 4
5/19 Thurs: Categorization and Summarization (slides)
Material covered: Kupiec et al. SIGIR95 paper, Chen-Bloomberg DimSum paper
Handouts: Solutions to Problem Set 4
5/24 Tues: e-books (Guest Lecturer: William C. Janssen) (slides)
Material covered: Reading on e-books
Handouts: Information about available code for affine transform
5/26 Thurs: Distributed Digital Libraries (Guest Lecturer: Fabio Crestani) (slides)
Material covered: Reading on distributed digital libraries
Handouts: Web links related to e-books
5/31 Tues: TBA (slides)
6/02 Thurs: Project Presentations
Due in class: Project reports