Showing papers in "Computer Vision and Image Understanding in 2006"

TL;DR: This survey reviews recent trends in video-based human capture and analysis, as well as discussing open problems for future research to achieve automatic visual analysis of human movement.

TL;DR: This survey focuses on recognition performed by matching models of the three-dimensional shape of the face, either alone or in combination with matching corresponding two-dimensional intensity images.

TL;DR: Results indicate that this MHV representation can be used to learn and recognize basic human action classes, independently of gender, body size and viewpoint.

TL;DR: Algorithms for recognizing human motion in monocular video sequences, based on discriminative conditional random field (CRF) and maximum entropy Markov models (MEMM) are presented, which outperform HMMs in classifying not only diverse human activities like walking, jumping.

TL;DR: The general camera placement problem is first defined with assumptions that are more consistent with the capabilities of real world cameras, and a solution to this problem is obtained via binary optimization over a discrete problem space.

TL;DR: An adaptation of Radon transform called R-transform is proposed, which is invariant to common geometrical transformations, and a binary shape is projected into the Radon space for different levels of the Chamfer distance transform.

TL;DR: A system for human behaviour recognition in video sequences that combines Bayesian networks and belief propagation, non-parametric sampling from a previously learned database of actions, and Hidden Markov Models which encode scene rules are used to smooth sequences of actions.

TL;DR: A novel color image enhancement method, which is named HVS Controlled Color Image Enhancement and Evaluation algorithm (HCCIEE algorithm), which is base on multiscale representation of pattern, luminance, and color processing in the HVS is proposed.

TL;DR: This study finds that landmarks and their geometry-based approach can account for variations of face expression and aging very well and can be used either in stand-alone mode or in conjunction with other approaches to reduce the search space a priori.

TL;DR: A new way of trating occlusions is presented by splitting segmented blobs based on morphological operators and a backward and forward graph representation which allows an increasing in the number of frames automatically tracked.

TL;DR: This paper evaluates the performance of several popular corner detectors using two newly defined criteria, consistency and accuracy, which show that the enhanced CSS corner detector performs better according to these criteria.

TL;DR: It is shown that a combination of multiple image cues helps the tracker to overcome ambiguous situations such as limbs touching or strong occlusions of body parts, and stochastic sampling makes SMD robust against local minima and lowers the computational costs as a small set of predicted image features is sufficient for optimization.

TL;DR: A dual-factor authentication methodology coined as S-Iris Encoding is proposed based on the iterated inner-products between the secret pseudo-random number and the iris feature, and with thresholding to produce a unique compact binary code per person.

TL;DR: An approach to 3D people tracking with learned motion models and deterministic optimization is explored, showing that it can learn and track cyclic motions such as walking and running, as well as acyclic motionssuch as a golf swing.

TL;DR: The purpose of this study is to investigate a new representation of a partition of an image domain into a fixed but arbitrary number of regions by explicit correspondence between the regions of segmentation and the regions defined by simple closed planar curves and their intersections.

TL;DR: It is proved that in the case of central parabolic systems and cameras with lens distortion the locus of the lifted points representing projections of world lines is a plane, and a new representation for the image plane of central systems is provided.

TL;DR: The estimated camera intrinsics model along with the cube-maps provides a calibration reference for images captured on the fly by the active pan-tilt-zoom camera under operation making the approach promising for active camera network calibration.

TL;DR: The proposed approach constitutes a principled unified probabilistic framework for low level scene analysis and understanding, showing several key features with respect to the state of the art methods, as it extracts information at the lowest possible level (using only pixel gray-level temporal behavior), and is unsupervised in nature.

TL;DR: A template-based approach to detecting human silhouettes in a specific walking pose using short sequences of 2D silhouettes obtained from motion capture data that helps distinguish actual people who move in a predictable way from static objects whose outlines roughly resemble those of humans.

TL;DR: A framework to simultaneously segment and track multiple body parts of interacting humans in the presence of mutual occlusion and shadow using an attribute relational graph based multi-target, multi-association tracking system and a coarse model of the human body.

TL;DR: This work introduces a generic structure-from-motion approach for this general imaging model, that allows to reconstruct scenes from calibrated images, possibly taken by cameras of different types (cross-camera scenarios), and proposes two approaches for obtaining optimal solutions using bundle adjustment.

TL;DR: This work proposes a generalized scale model which is spatially adaptive like other local morphometric models, and yet possesses the global spirit of multi-scale representations, and presents a variant of the generalized scale notion that is referred to as the generalized ball scale, which has superior noise resistance properties.

TL;DR: Results of single view tracking demonstrate that the exemplar-based models incorporating dynamics generalise to viewpoint invariant tracking of novel movements.

TL;DR: This work proposes a method called boundary matting, which represents each occlusion boundary as a 3D curve, and suggests that this method enables high-quality view synthesis with reduced matting artifacts.

TL;DR: In this article, a scale is defined as the radius of the largest hyperball contained in the same homogeneous region under a predefined condition of homogeneity of the image vector field.

TL;DR: An automated technique for locating previously unknown commercials by continuously monitoring broadcast television signals is presented and has consistently achieved over 93% accuracy identifying new commercials and non-commercials as they are broadcast.

TL;DR: A quantitative comparison between this method and a state-of-the-art shadow detection algorithm clearly indicates that this method is promising for delivering effective shadow detection performance under different illumination and brightness conditions.

TL;DR: A new color image segmentation scheme based on unsupervised pixel classification that works even when there is not a one-to-one correspondence between the clusters of color points in the color space and the regions in the image is presented.

TL;DR: This paper model the motions of independently moving cameras in the equations governing the epipolar geometry and derive a new relation which is referred to as the "temporal fundamental matrix" and calculates a matching score between two actions by evaluating the quality of the recovered geometry.

TL;DR: This work draws on spectral graph theory to derive a new algorithm for computing node correspondence in the presence of noise and occlusion, and demonstrates the approach on the domain of view-based 3-D object recognition.