************************************************************************ _______ IEEE SIGNAL PROCESSING SOCIETY /| /| / |ICIP / | 1995 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING /__|____/ | October 23-26, 1995 | |____|__| Hyatt Regency Crystal City | / | / Washington, D.C., USA | / 95 | / |/______|/ ON-LINE INFORMATION NEW Information: Session Titles, Tutorials, Registration, Hotels *************************************************************************** WWW address ftp site: ftp.ee.princeton.edu/pub/ICIP95 information: icip95@ieee.org See instructions at end of message. *************************************************************************** Mail to icip95@ieee.org or icip95@ccr-p.ida.org is processed by procmail. To ask a question or receive information, please include the appropriate keyword as the Subject of your message. (Note: one keyword per message) To ask a question or reach the Conference Managers use Subject: question - to ask a question ICIP - to reach the Conference Management. Note that their direct address is: mercer@conf-mgmt.com To receive an automatic reply on a specific subject use Subject: Information - Keywords, Program, Registration and Hotel information Keywords - This list of keywords. Program - Session Titles, Plenary Topics, and Tutorials. Registration- Information and form Hotel - Information and form Tutorials - Information and registration forms Call - The Call-for-Papers Style - The LaTeX style file for camera-ready papers Exhibits - The Exhibitor's Prospectus Note: If you make any inquiry about ICIP-95, you will be added to the ICIP mailing list and will receive all future information automatically as soon as it becomes available. If you wish to be removed from the mailing list, send a message containing the email address through which you received our mail with Subject: delete - Request to be deleted from the ICIP mailing list Conference Manager: Billene Mercer Conference Management Services mercer@conf-mgmt.com 2553 Texas Avenue South Suite C-283 College Station, TX 77840 Tel: 409-696-6596 Fax: 409-696-6653 ************************************************************************ ICIP-95 PLENARY SESSIONS Coding with Embedded Functionality - MPEG4 Content-Based Coding Dr. Cliff Reader Recent Developments in Medical Imaging and Image Processing Prof. Gabor Herman Image Processing and Analysis: What, Why, and How Prof. Azriel Rosenfeld *************************************************************************** SPECIAL SESSIONS Video on ATM Networks Digital Libraries Tele-Medicine Color and Printing Non-linear Dynamics in Image Processing Image Processing in Manufacturing and Materials Research Document Image Processing Signal and Image Processing Education *************************************************************************** TECHNICAL SESSIONS Monday, October 23, 1995 Morning Afternoon Multiresolution Non-Linear and Multi-Resolution Filters Architectures and Software Structure and Motion Estimation Vector Quantization I Applications I Real and Synthetic Aperture Radar Restoration/Enhancement Image Enhancement Morphology Motion Estimation I Motion-Based Video Coding Image Analysis Wavelet Compression Image Coding I Computed Imaging Tuesday, October 24, 1995 Image Restoration I Stereo and Shape Image Coding II Low Bit Rate Video Coding Parallel Algorithms Tomographic Theory and Algorithms Multi-Dimensional Filtering Medical Applications I Edge Detection Image Restoration II Motion Estimation II Video Compression I Multimedia Applications Facial Image Processing Image Models Object Recognition Wednesday, October 25, 1995 Image Segmentation Object Recognition Vector Quantization II Wireless and Video Rate Control Image Rendering and System Design Document Image Processing Hardware Implementation Medical Applications II Motion Segmentation and Estimation Video Compression II Applications II Fusion and Registration Image Segmentation Applications Fractal Coding Image Coding III Image Analysis II ************************************************************************ TUTORIAL DAY - Thursday, October 26, 1995 SESSION SA1: Markov Random Fields and Stochastic Image Models Time: 9:00 am - 12:00 noon Lecturer: Charles A. Bouman Purdue University In recent years, Markov random field (MRF) image models have been successfully applied in a wide variety of Bayesian image processing applications. The two primary advantages of MRF's are simple structure and accurate modeling of non-Gaussian image characteristics such as edges. This tutorial will review the history, current research and future directions of MRF's and other related stochastic image models in image processing applications. Specific topics will include causal AR models, simultaneous AR models, Ising and other discrete MRF models, critical temperature behavior, continuous MRF models, MRF parameter estimation, simulation and annealing, regularization and MAP estimation, and multiscale stochastic models. SESSION SA2: Splines, Atomic Spaces, and Wavelets Time: 9:00 am - 12:00 noon Lecturer: Akram Aldroubi National Institutes of Health Polynomial spline functions and atomic representations in general result in a characterization of continuous functions by discrete signals that is well adapted for digital processing. These representations provide a generalization of Shannon's sampling theory that is useful for implementing continuous operators using fast digital filtering algorithms. In particular, we will show how to design efficient algorithms for image interpolation, edge detection, noise reduction, and geometric image spline spaces is also well adapted for constructing fast multi-scale processing algorithms that start their computations at low resolution and refine the solution as they switch to finer scales. This is related to multigrid processing and is useful in solving problems with high computational cost such as 3-D image registration. The multiresolution properties of splines can also be used to construct wavelet bases or frames with various prescribed properties (e.g. optimal time-frequency localization). SESSION SP1: Medical Imaging: Modalities and Performance Time: 1:30 pm - 4:30 pm Lecturer: Oleh Tretiak Drexel University Image and signal processing has been a key element in the recent flourishing of medical imaging technologies. The presentation will review the modalities in medical imaging, both those in use at present and those under development. The review will include illustrative examples of x-ray, nuclear medicine, computer tomography, magnetic resonance imaging, and ultrasound. New development areas such as impedance tomography and functional MRI will be described. ROC evaluation of imaging instruments will be reviewed. SESSION SP2: Introduction to Partial Differential Equations in Image Processing Time: 1:30 pm - 4:30 pm Lecturer: Jean-Michel Morel CEREMADE, Universite de Paris 9 Dauphine, France By considering continuous image representations, most local filtering operations can be translated into differential operators. The simplest example is Gaussian smoothing which corresponds to the Laplacian. In the same way, we shall show that the median filter corresponds to the curvature of isophotes (iso-level curves), and the dilation operator to the modulus of the gradient. Iterating such local filters with a neighborhood that becomes infinitesimally small yields a partial differential equation (PDE). Thus, any iterative filtering process can be viewed as a partial differential evolution equation. What is gained by adopting the PDE methodology ? First, the elementary PDE formalism provides a new and independent means of classification of iterated filters; certain filters that appear to be different may have a similar structure that is only revealed by comparing their underlying PDEs. The formalism also provides a unification with "scale space" theories. Second, the discretization of PDE's yields new and more robust ways of implementing classical filters. Finally, some of the PDE's are new and not necessarily implementable through discrete filtering, although it is relatively easy to solve then numerically. The best example of this is the affine-invariant scale space representation which was discovered using PDE methods. The presentation will include many such image processing examples and provide a new viewpoint on several aspects of mathematical morphology. ***************************************************************************