2011-2012 Bulletin â PDF - SEAS Bulletin - Columbia University
2011-2012 Bulletin â PDF - SEAS Bulletin - Columbia University
2011-2012 Bulletin â PDF - SEAS Bulletin - Columbia University
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
118<br />
range imaging, mosaicing and superresolution.<br />
Classes are seminars with the instructor, guest<br />
speakers, and students presenting papers and<br />
discussing them.<br />
COMS E6733x or y 3D photography<br />
3 pts. Lect: 2.<br />
Prerequisite: Experience with at least one of the<br />
following topics: Computer graphics, computer<br />
vision, pixel processing, robotics or computeraided<br />
design, or permission of instructor.<br />
Programming proficiency in C, C++, or JAVA.<br />
3D Photography—the process of automatically<br />
creating 3D, texture-mapped models of objects<br />
in detail. Applications include robotics, medicine,<br />
graphics, virtual reality, entertainment and digital<br />
movies etc. Topics include 3D data acquisition<br />
devices, 3D modeling systems and algorithms<br />
to acquire, create, augment, manipulate, render,<br />
animate and physically build such models.<br />
COMS E6734y Computational photography<br />
3 pts. Lect: 3.<br />
Prerequisites: COMS W4160, W4731,<br />
or a working knowledge of photography<br />
are recommended. Students should have<br />
knowledge in any of three core areas: computer<br />
vision, computer graphics, or photography.<br />
Computational techniques are used to produce a<br />
new level of images and visual representations.<br />
Topics include HDR imaging, feature matching<br />
using RANSAC, image mosaics, image-based<br />
rendering, motion magnification, camera lens<br />
arrays, programmable lighting, face detection,<br />
single and multiview geometry, and more.<br />
COMS E6735y Visual databases<br />
3 pts. Lect: 3. Not offered in <strong>2011</strong>–<strong>2012</strong>.<br />
Prerequisites: COMS W3133, W3134, or W3137<br />
required. COMS W4731 and W4735 helpful but<br />
not required. Contact instructor if uncertain.<br />
The analysis and retrieval of large collections<br />
of image and video data, with emphasis on<br />
visual semantics, human psychology, and<br />
user interfaces. Low-level processing: features<br />
and similarity measures; shot detection; key<br />
frame selection; machine learning methods<br />
for classification. Middle-level processing:<br />
organizational rules for videos, including<br />
unedited (home, educational), semiedited<br />
(sports, talk shows), edited (news, drama);<br />
human memory limits; progressive refinement;<br />
visualization techniques; incorporation of audio<br />
and text. High-level processing: extraction of<br />
thematic structures; ontologies, semantic filters,<br />
and learning; personalization of summaries and<br />
interfaces; detection of pacing and emotions.<br />
Examples and demonstrations from commercial<br />
and research systems throughout. Substantial<br />
course project or term paper required.<br />
COMS E6737x or y Biometrics<br />
3 pts. Lect: 3.<br />
Prerequisite: Background at the sophomore<br />
level in computer science, engineering, or like<br />
discipline. Corequisites: None In this course we<br />
will explore the latest advances in biometrics as<br />
well as the machine learning techniques behind<br />
them. Students will learn how these technologies<br />
work and how they are sometimes defeated.<br />
Grading will be based on homework assignments<br />
and a final project. There will be no midterm<br />
or final exam. This course shares lectures with<br />
COMS W4737. Students taking COMS E6737<br />
are required to complete additional homework<br />
problems and undertake a more rigorous final<br />
project. Students will only be allowed to earn credit<br />
for COMS W4737 or COMS E6737 but not both.<br />
CSEE E6824y Parallel computer architecture<br />
3 pts. Lect: 2.<br />
Prerequisite: CSEE W4824. Parallel computer<br />
principles, machine organization and design of<br />
parallel systems including parallelism detection<br />
methods, synchronization, data coherence and<br />
interconnection networks. Performance analysis<br />
and special purpose parallel machines.<br />
CSEE E6831y Sequential logic circuits<br />
3 pts. Lect: 3. Not offered in <strong>2011</strong>–<strong>2012</strong>.<br />
Prerequisite: CSEE W3827 or any introduction<br />
to logic circuits. Generation and manipulation<br />
of flow table descriptions to asynchronous<br />
sequential functions. Coding of flow tables to<br />
satisfy various design criteria. Delays, races,<br />
hazards, metastability. Analysis of latches to<br />
determine key parameters. Bounds of input<br />
rates. Clocking schemes for synchronous<br />
systems. Synthesis of self-timed systems using<br />
4-phase or 2-phase handshakes.<br />
CSEE E6832x or y Topics in logic design<br />
theory<br />
3 pts. Lect: 3.<br />
Prerequisite: CSEE W3827 or any introduction to<br />
logic circuits. A list of topics for each offering of<br />
the course is available in the department office<br />
one month before registration. May be taken<br />
more than once if topics are different Iterative<br />
logic circuits applied to pattern recognition. Finite<br />
state machines; alternative representations,<br />
information loss, linear circuits, structure theory.<br />
Reliability and testability of digital systems.<br />
CSEE E6847y Distributed embedded systems<br />
3 pts. Lect: 2. Not offered in <strong>2011</strong>–<strong>2012</strong>.<br />
Prerequisite: Any COMS W411X, CSEE W48XX,<br />
or ELEN E43XX course, or permission of<br />
instructor. An interdisciplinary graduate-level<br />
seminar on the design of distributed embedded<br />
systems. System robustness in the presence<br />
of highly variable communication delays and<br />
heterogeneous component behaviors. The<br />
study of the enabling technologies (VLSI<br />
circuits, communication protocols, embedded<br />
processors, RTOSs), models of computation,<br />
and design methods. The analysis of modern<br />
domain-specific applications including on-chip<br />
micro-networks, multiprocessor systems, faulttolerant<br />
architectures, and robust deployment<br />
of embedded software. Research challenges<br />
such as design complexity, reliability, scalability,<br />
safety, and security. The course requires<br />
substantial reading, class participation and a<br />
research project.<br />
CSEE E6861y Computer-aided design of<br />
digital systems<br />
3 pts. Lect: 2.<br />
Prerequisites: (i) one semester of advanced<br />
digital logic (CSEE W4823 or equivalent,<br />
or instructor’s permission); and (ii) a basic<br />
course in data structures and algorithms<br />
COMS W3133, 3134, 3137, 3139, or 3157, or<br />
equivalent, and familiarity with programming.<br />
Introduction to modern digital CAD synthesis<br />
and optimization techniques. Topics include:<br />
modern digital system design (high-level<br />
synthesis, register-transfer level modeling,<br />
algorithmic state machines, optimal scheduling<br />
algorithms, resource allocation and binding,<br />
retiming), controller synthesis and optimization,<br />
exact and heuristic two-level logic minimization,<br />
advanced multilevel logic optimization, optimal<br />
technology mapping to library cells (for delay,<br />
power and area minimization), advanced<br />
data structures (binary decision diagrams),<br />
SAT solvers and their applications, static<br />
timing analysis, and introduction to testability.<br />
Includes hands-on small design projects using<br />
and creating CAD tools.<br />
EECS E6870x or y Speech recognition<br />
3 pts. Lect: 3.<br />
Prerequisites: Basic probability and statistics.<br />
Theory and practice of contemporary<br />
automatic speech recognition. Gaussian<br />
mixture distributions, hidden Markov models,<br />
pronunciation modeling, decision trees, finitestate<br />
transducers, and language modeling.<br />
Selected advanced topics will be covered in<br />
more depth.<br />
COMS E6900x and y Tutorial in computer<br />
science<br />
1–3 pts.<br />
Prerequisite: Instructor’s permission. A reading<br />
course in an advanced topic for a small number<br />
of students, under faculty supervision.<br />
COMS E6901x Projects in computer science<br />
1–12 pts.<br />
Prerequisite: Instructor’s permission. Software<br />
or hardware projects in computer science.<br />
Before registering, the student must submit a<br />
written proposal to the instructor for review.<br />
The proposal should give a brief outline of the<br />
project, estimated schedule of completion, and<br />
computer resources needed. Oral and written<br />
reports are required. May be taken over more<br />
than one semester, in which case the grade<br />
will be deferred until all 12 points have been<br />
completed. No more than 12 points of COMS<br />
E6901 may be taken. Consult the department for<br />
section assignment.<br />
engineering <strong>2011</strong>–<strong>2012</strong>