============== CVPR Tutorials ============== Sunday, June 19 1:00 - 5:00 pm Shape and Image Representation for Recognition: Geometric Deformations, Scale, and Nonlinear Diffusion Instructor: Benjamin B. Kimia, Brown University A general framework for representing shape and images based on differential geometeric deformations of them is represented for the purpose of object recognition. This framework embeds notions of skeleltons but includes classification of points AND a notion of significance, doing away with classical noise problems there. Another topic is mathematical morphology which is also a special case under this framework. We will present numerical algorithms for very robust decompositions of industrial and biomedical shapes into functional parts. In addition, the deformable contours are exellent alternatives for snakes. We will show how this can be generalized to three-dimensional deformations and the resulting scale-space for surface estimates and images. Sunday, June 19 1:00 - 5:00 pm Applications of Neural Networks in Signal and Image Processing Instructor: Nasser M. Nasrabadi, SUNY at Buffalo This course is designed for engineers interested in exploring the applications of ANNs to signal and image processing. Topics will include human visual system, visual cortex, study of neurons receptive fields, multi-layer networks, time delay neural network, radial basis function, Kohonen feature maps, Hopfield neural network, simulated annealing for global optimization techniques. Applications of ANNs includes stereo vision, motion analysis, surface interpolation, image restoration, image segmentation, object recognition, optical character recognition, recognition of handwritten English words, image compression, non-linear predictions and robotics applications. Monday, June 20 8:00 am - 12 noon Biomedical Image Processing Instructor: Raj Acharia, SUNY at Buffalo In this tutorial, we will briefly review the various Medical Imaging Modalities such as MRI, CT and PET. We will also provide an overview of the key problems in Biomedical Image Analysis. We will review problems in Mul tiresolution Texture Analysis for Bone Tissue Characterization, Segmentation, Motion Analysis, Physiological/Functional Models and Multimodality Imaging. Monday, June 20 8:00 am - 12 noon Document Understanding Instructor: Sargur N. Srihari, CEDAR, SUNY at Buffalo Monday, 6/20/94 This tutorial will cover principles of document analysis, recognition and understanding. Both handwriting and machine-print recognition will be reviewed. Transition of research algorithms into real-time systems will be described. Monday, June 20 8:00 am - 12 noon Three-dimensional object recognition Instructor: Patrick Flynn, Washington State University This tutorial will explore each of the logical modules of model-based 3D object recognition systems. We will briefly survey the state of the art in 3D imaging sensors, explore image features commonly used for 3D vision and methods for their extraction from imagery, address 3D modeling techniques, and present a number of poular strategies for matching image features to stored models. Monday, June 20 1:00 - 5:00 pm Visual Databases and Multimedia Instructor: Ramesh Jain, University of California at San Diego In many applications, a large volume of image and video data must be, organized to allow efficient retrieval of information. The queries in such, databases will use alphanumeric and multimedia entities. In this tutorial, we will discuss techniques for database management systems that can answer content queries in databases containing video, graphics, images, and other non-alphanumeric data. Monday, June 20 1:00 - 5:00 pm Statistical Pattern Recognition Instructor: Anil Jain, Michigan State University Pattern recognition techniques are computer-based procedures for automatically classifying objects and making decisions. Commercial pattern recognition systems exist for printed text, blood cells, fingerprints, voice recognition, and word recognition. Most industrial machine vision systems employ pattern recognition to identify objects for sorting, inspection and assembly. The objective of this tutorial is to introduce fundamental methods of statistical pattern recognition with examples from several application areas. Monday, June 20 1:00 - 5:00 pm Mathematical Morphology Instructor: Robert Haralick, University of Washington This tutorial will visually illustrate the basic principles of Mathematical Morphology. Binary and gray scale dilation, erosion, opening, and closing will be covered as well as the opening and closing transforms which use recursive morphology. Applications in shape extraction, document image analysis, and noise cleaning will be shown.