AC52021: Vision and Perception

PEOPLE:

TIMETABLE:

Lectures will take place in two time slots every week as follows
Monday 11:00-13:00 in Dalhousie 2S11(B) and
Friday 10:00-12:00 in Dalhousie 2F10(B)
Lab sessions will be held in QM Lab 2 on
Wednesday 9:00-11:00

full MSC-ICS timetable is here

Detailed Schedule (locations as above unless otherwise noted):

STRUCTURE

The course will consist of 7 topics (see below). For each topic, there will be an in-class lecture (usually 1.5 hours), followed by a discussion of a research paper (also in class, usually 2 hours).

The discussions will be structured as follows. One person will be chosen each week to present the paper, by giving a short (10-15 min) overview. The other participants are required to bring one or more exam-style discussion points/questions, or arguments against the paper. These arguments should be written down before the discussion, and the students should be prepared to discuss their arguments with the group. The written argument should be at least two sentences, preferably a short paragraph. Participation marks will be awarded for a written argument, followed by a comprehensible discussion of the argument. Examples of arguments may be

Each of these examples should be followed by a brief explanation or a few discussion points. See more discussion points on the SIFT paper. You don't have to actually know for certain that your argument is correct or what the answer is, just be prepared to discuss it. Handed-in arguments will also form the basis for exam questions.

The course content will be delivered partially as lectures, and partially as group discussion of computer vision papers. Each week, students are expected to read the discussion paper, prepare one or more arguments, and participate in the discussion. Further their will be two short projects to be completed and write-ups to be handed in.

Assessment

READINGS:

Primary Texts:

RESOURCES:

COURSE OUTLINE

Topic 1: Features (Hoey)


Lecture Material:

Discussion Reading:

David G. Lowe, Object recognition from local scale-invariant features,
International Conference on Computer Vision, Corfu, Greece (September 1999), pp. 1150-1157.


Related Readings:

Project #1:

Scale Invariant Feature Transform. See project description page.
Due Date: Februrary 23rd (in class)

Topic 2: Optical Flow (Trucco)

Lecture Material:

Discussion Reading (Fri Feb 6th, 2009):

B K P Horn and B G Schunk: Determining optic flow. Artificial Intelligence Journal, 1981. (take a look at Ballard and Brown's overview first; see link in "Lecture Material")

Related Material:

optical flow test sequences
John Barron and Steven Beauchemin, Performance of Optical Flow Techniques. IJCV 12(1):43-77, 1994.
get the pdf here

Topic 3: Segmentation (Trucco)

Lecture Material:

Discussion Reading (Fri Feb 13th, 2009):

D. Comaniciu, P. Meer: Mean Shift Analysis and Applications, IEEE Int. Conf. Computer Vision (ICCV'99), Kerkyra, Greece, 1197-1203, 1999

Related Material:

A review and comparison of various segmentation techniques (ECCV 2002) by Freixenet et al.
Dorin Comaniciu and Peter Meer: Mean Shift: A Robust Approach toward Feature Space Analysis IEEE PAMI vol 24, No 5, May 2002. A longer version of the discussion reading.
Dorin Comaniciu home page
C/matlab code for doing normalized cuts, and a fancier version

Topic 4: Contours and Shape (McKenna)


Lecture Material:

See the share.

Discussion Reading:

Cootes T F, Taylor C J, Cooper D H, Graham J, Active shape models - their training and application, Computer Vision and Image Understanding 61(1), 38-59, 1995

Related Readings:

Kass M, Witkin A, Terzopoulos D, Snakes: active contour models, International Journal of Computer Vision 1(4), 321-331, 1988
CVonline: Snakes, Active Contours

Topic 5: Object Recognition (Hoey)

Lecture Material:

Discussion Reading:

Viola, Paul, and Jones, M. Rapid object detection using a boosted cascade of simple features. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Dec. 2001).

Project #2:

Face Detection with a Boosted Feature Detector. See project description page.
Due Date: March 30th (in class)

Topic 6: Tracking (McKenna)

Lecture Material:

Discussion Reading:

Isard M and Blake A, Contour tracking by stochastic propagation of conditional density, European Conference on Computer Vision, Cambridge, 343-356, 1996

Related Readings:

Vermaak J, Gangnet M, Blake A, Perez P, Sequential Monte Carlo Fusion of Sound and Vision for Speaker Tracking, International Conference on Computer Vision 2001
Nait Charif H and McKenna S J, Tracking the Activity of Participants in a Meeting, Machine Vision and Applications 17(2):83-93, May 2006


Topic 7: Action/Event Recognition (Hoey)

Lecture Material:

Discussion Reading 1:

Alexei A. Efros, Alexander Berg, Greg Mori, Jitendra Malik. Recognizing Action at a Distance. In Proceedings of International Conference on Computer Vision, 2003

ICCV talk

Related Readings:


Sourabh A. Niyogi, Edward H. Adelson. Analyzing and Recognizing Walking Figures in XYT Proc. IEEE Conf. Computer Vision and Pattern Recognition, pp.469-474, 1994
James J. Little and Jeffrey E. Boyd Recognising People by their Gait: The Shape of Motion, Videre, Volume 1, Number 2, Winter 1988
Thad E. Starner and Alex P. Pentland Visual Recognition of American Sign Language using Hidden Markov Models Proc. IEEE Conf on Automatic Face and Gesture Recognition, Zurich, 1995
Kevin P. Murphy and Mark A. Paskin Linear Time Inference in Hierarchical Hidden Markov Models, Proceedings of Neural Information Processing Systems (NIPS), Vancouver, BC, 2001
Nuria Oliver and Eric Horvitz A Comparison of HMMs and Dynamic Bayesian Networks for Recognizing Office Activities Proceedings of Int. Conf. on User Modeling (UM'05). Edinburgh, UK. Jul 2005.
Eric Horvitz, Jack Breese, David Heckerman, David Hovel, Koos Rommelse The Lumiere Project: Bayesian User Modeling for Inferring the Goals and Needs of Software Users. In Procedings of International Conference on Uncertainty in Artificial Intelligence, Madison, WI, 1998.