Tutorial
Date: December 15, 2004.
Venue: Indian Statistical Institute, Kolkata.
Limited number of participants can be accommodated.
Registration Fee: Rs. 2,000/- or US $ 100.
Tuorial Session
Morning Session
9:00 - 10:00 Registration 9:45 - 10:00 Introduction by S. K. Pal (Tutorial Chair) 10:00 - 11:30 Andrew Zisserman 11:30 - 11:45 Tea break 11:45 - 13:15 Baba C. Vemuri 13:15 - 14:15
Lunch
Afternoon Session 14:15 - 15:45 Larry Davis 15:45 - 16:00 Tea break 16:00 - 17:30 J. K. Aggarwal
Download Registration Form (PDF)
Mode of payment:
Mail the Registration form along with the payment to the following address:
Secretariat,
ICVGIP 2004
Electronics
and Communication Sciences Unit
INDIAN
STATISTICAL INSTITUTE
203
B. T. Road, Calcutta 700108, INDIA
Email:
icvgip04@isical.ac.in
Tel:
+91 33 25752915, 25752913, 25753108
Fax:
+91 33 25773035, +91 33 25776680.
|
University of Maryland, USA |
|
| Topic: | Detection and tracking for surveillance |
| Abstract: | This
tutorial will cover problems related to detection and tracking from
surveillance video. For detection,
I will cover background subtraction for fixed and pan/tilt/zoom cameras,
shape-based detection using machine learning techniques, and the
use of periodic motion models to discriminate people from other moving
objects in video. For tracking, I will discuss discuss basic
representation and control problems in tracking objects both continuously
and with time gaps in coverage. |
|
University of Texas, USA |
|
| Topic: | Human Motion: Actions and Interactions |
| Abstract: | |
|
University of Oxford, UK |
|
| Topic: | Trainable Visual Models for Object Class Recognition |
| Abstract: | Recognizing object classes, such as cars, planes or elephants, in an image or a video remains one of the most challenging problems in Computer Vision. However, recently a number of successes have been achieved by using ideas and algorithms from statistical learning theory, where visual models are trained using positive and negative examples of the class. |
|
University of Florida, USA |
|
| Topic: | Image Registration with Applications to Medical Imaging |
| Abstract: |
Image registration is ubiquitous in Computer Vision, Medical Imaging and other fields of Science and Engineering. Image registration maybe defined as, given a pair of images possibly acquired from different view points under varying image acquisition parameters or from two different sensors, to estimate the unknown coordinate transformation that would align the images. The coordinate transformation may be linear (e.g., rigid, affine etc.) or nonlinear (e.g., polynomial etc.). Several well known approaches (e.g., the SSD method, the Mutual Information method etc.) to solve this problem will first be reviewed and following this, I will present a relatively new method that involves a novel measure of information in a random variable based on its cumulative distribution and dubbed the cumulative residual entropy (CRE). This measure parallels the well known Shannon entropy but has the following advantages: (1) it is more general than the Shannon Entropy as its definition is valid in the discrete and continuous domains, (2) it possess more general mathematical properties and (3) it can be easily computed from sample data and these computations asymptotically converge to the true values. Based on CRE, I will define the cross-CRE (CCRE) between two random variables, and apply it to solve the image alignment problem for parameterized (3D rigid and affine) transformations. The key strengths of the CCRE over using the mutual information (based on Shannon’s entropy) are that the former has significantly larger tolerance to noise and a much larger convergence range over the field of parameterized transformations. We demonstrate these strengths via experiments on synthesized and real image data. |