Polytechnic
University, Electrical Engineering
EL612 ----
Video Processing, Spring 2001
Course
Description: This
course introduces fundamental theory and techniques for efficient representation
and processing of video signals in digital format. Topics to be covered
include: Fourier analysis of video signals, spatial and temporal frequency
responses of the human visual system, sampling and sampling rate conversion
of video signals, motion estimation, basic video compression techniques,
and video communication standards. A mini-project is required.
Prerequisites:EL512,
E630, and graduate status. Instructor approval required for senior students.
This course can be used to form a two-course sequence with EL512.
Instructor:Professor
Yao Wang, LC259, (718)-260-3469, LC256, yao@vision.poly.edu,
http://www.ee.poly.edu/~yao
Course
Schedule: Spring
2001, Monday 6:00- 8:15 PM
Office
Hour: Monday 4:00-5:00, Wed. 11:00-12:00
Text
Book: Y.
Wang, J. Ostermann, and Y.Q.Zhang, Digital Video Processing, To
be published by Prentice Hall, 2001. Copies of manuscript will be provided
to students.
Recommended
Readings:A. N. Netravali and B. G. Haskell, Digital Pictures
– Representation, Compression and Standards, 2nd
ed. Plenum Press, 1995.
Grading
Policy: Homework
70%, Project (including project presentation and report) 30%. No exams.
Website:http://www.ee.poly.edu/~yao/EL612
Tentative
Course Outline
-
Basics of analog and digital
video basics: color video formation and specification, analog TV system
(NTSC/PAL/SECAM), video raster, digital video formats (CCIR601, HDTV, etc.)
(1 lecture)
-
Frequency domain analysis of
video signals, spatial and temporal frequency response of the human visual
system. (1 lecture)
-
Video Sampling: lattice sampling
theorem for multidimensional signal sampling; video signal sampling (raster
scan, interlaced vs progressive scan), characteristics of common TV camera
and monitor. (1 lecture)
-
Video sampling rate conversion:
conversion of signals sampled over different lattices, sampling rate conversion
of video signals (interlaced to/from progressive scan, low definition to/from
high definition, frame rate conversion) (1 lecture)
-
Scene, camera, and motion modeling,
3D motion and projected 2D motion, models for typical camera/object motions.
(1 lecture)
-
2D motion estimation: optical
flow equation, different motion estimation methods (pel-based, block-based,
mesh-based, global motion estimation, multi--resolution approach) (2 lecture)
-
Basic compression techniques:
information bounds for lossless and lossy source coding, entropy coding,
scalar/vector quantization (1 lecture)
-
Waveform-based coding: predictive
coding, transform coding, motion compensated frame prediction and interpolation,
block-based hybrid video coding (2 lecture)
-
Video segmentation and 3D motion
estimation (1 Lecture)
-
Content-based video compression
(1 lecture)
-
Video compression standards
(H.261 and H.263, MPEG1, MPEG2, MPEG4). (2 lecture)