Showing papers by "Azriel Rosenfeld published in 2000"
TL;DR: A computer vision system for tracking multiple people in relatively unconstrained environments is described and should provide a useful mechanism for bootstrapping and reinitialization of tracking using more specific but less robust human models.
755 citations
03 Sep 2000
TL;DR: Confidence maps gray-scale images whose intensity is a function of confidence that a pixel has changed - to fuse intermediate results and represent the results of background subtraction.
Abstract: A new method of finding people in video images is presented. The detection is based on a novel background modeling and subtraction approach which uses both color and edge information. We introduce confidence maps gray-scale images whose intensity is a function of confidence that a pixel has changed - to fuse intermediate results and represent the results of background subtraction. The latter is used to delineate a person's body by guiding contour collection to segment the person from the background. The method is tolerant to scene clutter, slow illumination changes, and camera noise, and runs in near real time on a standard platform.
304 citations
TL;DR: This thesis approaches the off-line problem by establishing a local correspondence between a model and a questioned signature, and performs skilled forgery detection by examining the writer-dependent information embedded at the sub-stroke level and trying to capture unballistic motion and tremor information in each stroke segment, rather than as global statistics.
Abstract: Signatures may be stylish or unconventional and have many personal characteristics that are challenging to reproduce by anyone other then the original author. For this reason, signatures are used and accepted as proof of authorship or consent on personal checks, credit purchases and legal documents. Currently signatures are verified only informally in many environments, but the rapid development of computer technology has stimulated great interest in research on automated signature verification and forgery detection.
In this thesis, we focus on forgery detection of off-line signatures. Although a great deal of work has been done on off-line signature verification over the past two decades, the field is not as mature as on-line verification. Temporal information used in on-line verification is not available off-line and the subtle details necessary for off-line verification are embedded at the stroke level and are hard to recover robustly. We approach the off-line problem by establishing a local correspondence between a model and a questioned signature. The questioned signature is segmented into consecutive stroke segments that are matched to the stroke segments of the model. The cost of the match is determined by comparing a set of geometric properties of the corresponding sub-strokes and computing a weighted sum of the property value differences. The least invariant features of the least invariant sub-strokes are given the biggest weight, thus emphasizing features that are highly writer-dependent.
Random forgeries are detected when a good correspondence cannot be found, i.e., the process of making the correspondence yields a high cost. Many simple forgeries can also be identified in this way. The threshold for making these decisions is determined by a Gaussian statistical model.
Using the local correspondence between the model and a questioned signature, we perform skilled forgery detection by examining the writer-dependent information embedded at the sub-stroke level and trying to capture unballistic motion and tremor information in each stroke segment, rather than as global statistics. Experiments on random, simple and skilled forgery detection are presented.
57 citations
TL;DR: A bibliography of nearly 1700 references related to computer vision and image analysis, arranged by subject matter is presented, including computational techniques; feature detection and segmentation; image and scene analysis; and motion.
39 citations
TL;DR: This paper shows how to detect families of parallel lines in 3-D space at a moderate computational cost by using a (2+2)-D Hough space by finding peaks in the 2-D slope parameter space.
38 citations
TL;DR: Methods of determining, for an SN partial tiling P, whether a tile P∈ P ′ is simple, and if not, of counting the numbers of components and holes in N ∗ (P, P ) .
22 citations
TL;DR: “What being walks sometimes on two feet, sometimes on three, and sometimes on four, and is weakest when it has the most?” —The Sphinx's Riddle
Abstract: “What being walks sometimes on two feet, sometimes on three, and sometimes on four, and is weakest when it has the most?” —The Sphinx's Riddle
Pattern recognition is one of the most important functionalities for intelligent behavior and is displayed by both biological and artificial systems Pattern recognition systems have four major components: data acquisition and collection, feature extraction and representation, similarity detection and pattern classifier design, and performance evaluation In addition, pattern recognition systems are successful to the extent that they can continuously adapt and learn from examples; the underlying framework for building such systems is predictive learning The pattern recognition problem is a special case of the more general problem of statistical regression; it seeks an approximating function that minimizes the probability of misclassification In this framework, data representation requires the specification of a basis set of approximating functions Classification requires an inductive principle to design and model the classifier and an optimization or learning procedure for classifier parameter estimation Pattern recognition also involves categorization: making sense of patterns not previously seen The sections of this paper deal with the categorization and functional approximation problems; the four components of a pattern recognition system; and trends in predictive learning, feature selection using “natural” bases, and the use of mixtures of experts in classification © 2000 John Wiley & Sons, Inc Int J Imaging Syst Technol 11, 101–116, 2000
20 citations
TL;DR: Using these techniques, a viewer can browse a video at the level of summarization that suits his patience level, and applications include the creation of a smart fast-forward function for digital VCRs, and the automatic creation of short summaries that can be used as previews before videos are downloaded from the web.
Abstract: A video can be mapped into a multidimensional signal in a non-Euclidean space, in a way that translates the more predictable passages of the video into linear sections of the signal. These linear sections can be filtered out by techniques similar to those used for simplifying planar curves. Different degrees of simplification can be selected. We have refined such a technique so that it can make use of probabilistic distances between statistical image models of the video frames. These models are obtained by applying hidden Markov model techniques to random walks across the images. Using our techniques, a viewer can browse a video at the level of summarization that suits his patience level. Applications include the creation of a smart fast-forward function for digital VCRs, and the automatic creation of short summaries that can be used as previews before videos are downloaded from the web.
20 citations
01 Sep 2000-Graphical Models \/graphical Models and Image Processing \/computer Vision, Graphics, and Image Processing
TL;DR: It can be shown that if each voxel P of S has only finitely many neighbors (voxels of S that intersect P), and if any nonempty intersection of neighbors of P intersects P, then the neighborhood N(P) of every voxels P is simply connected and without cavities.
Abstract: Classical digital geometry deals with sets of cubical voxels (or square pixels) that can share faces, edges, or vertices, but basic parts of digital geometry can be generalized to sets S of convex voxels (or pixels) that can have arbitrary intersections. In particular, it can be shown that if each voxel P of S has only finitely many neighbors (voxels of S that intersect P), and if any nonempty intersection of neighbors of P intersects P, then the neighborhood N(P) of every voxel P is simply connected and without cavities, and if the topology of N(P) does not change when P is deleted (i.e., P is a “simple” voxel), then deletion of P does not change the topology of S.
13 citations
TL;DR: If G is a simply connected set of lattice points with graph structure defined by 4-neighbor adjacency, it is shown that the center of G is either a 2×2 square block, a diagonal staircase, or a (dotted) diagonal line with no gaps.
01 Sep 2000
TL;DR: If the authors histogram the normal flow vectors in images of a scene viewed by a moving observer, the time-varying histogram can be used to derive qualitative information about the observer's motion-for example, whether it is (primarily) translational or rotational.
Abstract: If we histogram the normal flow vectors in images of a scene viewed by a moving observer, we can use the time-varying histogram to derive qualitative information about the observer's motion-for example, whether it is (primarily) translational or rotational, and whether the direction of translation or axis of rotation is (roughly) parallel or perpendicular to the camera axis. This is illustrated using flow histogram obtained from a variety of real image sequences.
10 Sep 2000
TL;DR: This work derives an optimal smoothing filter, which minimizes both the noise power and the mean squared error between the input and the filter output, and defines an operator for shape detection by extending the DODE filter along the shape's boundary contour.
Abstract: We present a new approach for accurate detection of two-dimensional shapes. We first derive an optimal smoothing filter, which minimizes both the noise power and the mean squared error between the input and the filter output. This operator is found to be a derivative of the double exponential (DODE) function. We define an operator for shape detection by extending the DODE filter along the shape's boundary contour. We find that this filtering scheme is equivalent to integrating gradients along the hypothetical shape boundary, but our method turns out to be more robust than conventional edge detection followed by edge grouping. This approach also provides a tool for a systematic analysis of edge-based shape detection. We investigate how the error is propagated by the shape geometry. This enables us to predict both its localization and detection performance. Application to vehicle detection in aerial images and human facial feature detection are provided.
TL;DR: It should be possible to semiautomatically compile bibliographies of the current literature by collecting references from existing databases or publishers' web pages, standardizing their format, and classifying them using key words, but to ensure that the resulting bibliography is of good stylistic quality, manual postediting will be necessary.
13 Sep 2000
TL;DR: This work presents a polygon simplification method that works in multidimensional spaces or even in semi-metric spaces that need not be vector spaces, and requires only that a (semi-)distance between pairs of points be defined that needs not satisfy the triangle inequality.
Abstract: We present a polygon simplification method that works in multidimensional spaces or even in semi-metric spaces that need not be vector spaces. We require only that a (semi-)distance between pairs of points be defined that need not satisfy the triangle inequality.
01 Jan 2000
TL;DR: This chapter considers the topology of two-dimensional fuzzy digital pictures, and discusses topology-preserving deformations of these fuzzy pictures that generalize deformationsof two-valued digital pictures.
Abstract: This chapter considers the topology of two-dimensional fuzzy digital pictures, and discusses topology-preserving deformations of these fuzzy pictures that generalize deformations of two-valued digital pictures. It is shown that an arbitrary fuzzy picture can be transformed by this type of deformation into a coherent fuzzy picture. As another application, it is shown that the genus of a fuzzy picture is a linear combination of local property values.
01 Jan 2000
TL;DR: This dissertation defines a general framework for describing the physical and logical structure of documents, and describes a general system for document image retrieval that is able to make use of document structure.
Abstract: Title of Dissertation: The Roles of Document Structure in Document Image Retrieval and Classification Christian Kwang-Un Shin, Doctor of Philosophy, 2000 Dissertation directed by: Professor Azriel Rosenfeld Department of Computer Science Current document management and database systems provide text search and retrieval capabilities, but generally lack the ability to utilize the documents’ logical and physical structures. This dissertation defines a general framework for describing the physical and logical structure of documents, and describes a general system for document image retrieval that is able to make use of document structure. It discusses the use of structural similarity for retrieval; it defines a measure of structural similarity between document images based on content area overlap, and also compares similarity ratings based on this measure with human relevance judgments. Finally, it investigates document type classification using features related to physical layout structure, and using both decision-tree and self-organizing map classifiers; in these experiments too, ground truth was provided by human judgments. The Roles of Document Structure in Document Image Retrieval and Classification by Christian Kwang-Un Shin Dissertation submitted to the Faculty of the Graduate School of The University of Maryland, College Park in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2000 Advisory Committee: Professor Azriel Rosenfeld, Chairman/Advisor Doctor David S. Doermann Professor Larry S. Davis Doctor Daniel DeMenthon Professor Kyu Yong Choi, Dean’s Representative c Copyright by Christian Kwang-Un Shin 2000