Showing papers by "Azriel Rosenfeld published in 2001"

Classification of document pages using structure-based features

Abstract: Searching for documents by their type or genre is a natural way to enhance the effectiveness of document retrieval. The layout of a document contains a significant amount of information that can be used to classify it by type in the absence of domain-specific models. Our approach to classification is based on “visual similarity” of layout structure and is implemented by building a supervised classifier, given examples of each class. We use image features such as percentages of text and non-text (graphics, images, tables, and rulings) content regions, column structures, relative point sizes of fonts, density of content area, and statistics of features of connected components which can be derived without class knowledge. In order to obtain class labels for training samples, we conducted a study where subjects ranked document pages with respect to their resemblance to representative page images. Class labels can also be assigned based on known document types, or can be defined by the user. We implemented our classification scheme using decision tree classifiers and self-organizing maps.

3D object tracking using shape-encoded particle propagation

Abstract: We present a comprehensive treatment of 3D object tracking by posing it as a nonlinear state estimation problem. The measurements are derived using the outputs of shape-encoded filters. The nonlinear state estimation is performed by solving the Zakai equation, and we use the branching particle propagation method for computing the solution. The unnormalized conditional density for the solution to the Zakai equation is realized by the weight of the particle. We first sample a set of particles approximating the initial distribution of the state vector conditioned on the observations, where each particle encodes the set of geometric parameters of the object. The weight of the particle represents geometric and temporal fit, which is computed bottom-up from the raw image using a shape-encoded filter. The particles branch so that the mean number of offspring is proportional to the weight. Time update is handled by employing a second-order motion model, combined with local stochastic search to minimize the prediction error. The prediction adjustment suggested by system identification theory is empirically verified to contribute to global stability. The amount of diffusion is effectively adjusted using a Kalman updating of the covariance matrix. WE have successfully applied this method to human head tracking, where we estimate head motion and compute structure using simple head and facial feature models.

Tracking of human activities using shape-encoded particle propagation

Abstract: We present an approach to tracking human activities in a monocular video. We model the human body by decomposing it into torso and limbs and use simple 3D shapes to approximate them. The limb motions are parametrized by the relative joint angles. The problems of motion tracking and estimation are posed as nonlinear state estimation problems. The measurements are computed using the outputs of 3D shape-encoded filters which extract the boundary gradient information of the body image. The uncertainties of body pose are propagated by a branching particle system. We first sample a set of particles approximating the initial distribution of the state vector conditioned on observations, where each particle encodes the body pose. The posterior density is realized by the weight of the particle, where the weight represents geometric and temporal fit, and computed bottom-up from the raw image using a shape-encoded filter. The particles branch so that the mean number of offspring is proportional to the weight. Applications to both synthetic and real video sequences show the effectiveness of this approach.

Extraction of key frames from videos by polygon simplification

Abstract: We apply the polygon simplification method to automatically obtain the most relevant key frames. First a video sequence is mapped to a polyline in R/sup 37/. By simplifying this polyline, we obtain a summarization (i.e., a small set of the most relevant frames) that is representative of the whole video sequence. The degree of the simplification is either determined automatically or selected by the user.

Point-to-Line Mappings and Hough Transforms

Abstract: Nearly 40 years ago Hough showed how a point-to-line mapping that takes collinear points into concurrent lines can be used to detect collinear sets of points, since such points give rise to peaks where the corresponding lines intersect. Over the past 30 years many variations and generalizations of Hough's idea have been proposed, Hough's mapping was linear, but most or all of the mappings studied since then have been nonlinear, and take collinear points into concurrent curves rather than concurrent lines; little or no work has appeared in the pattern recognition literature on mappings that take points into lines.This paper deals with point-to-line mappings in the real projective plane. (We work in the projective plane to avoid the need to deal with special cases in which collinear points are mapped into parallel, rather than concurrent, lines.) We review basic properties of linear point-to-point mappings (collineations) and point-to-line mappings (correlations), and show that any one-to-one point-to-line mapping that takes collinear points into concurrent lines must in fact be linear. We describe ways in which the matrices of such mappings can be put into canonical form, and show that Hough's mapping is only one of a large class of inequivalent mappings. We show that any one-to-one point-to-line mapping that has an incidence-symmetry property must be linear and must have a symmetric matrix which has a diagonal canonical form. We establish useful geometric properties of such mappings, especially in cases where their matrices define nonempty conics.