scispace - formally typeset
Search or ask a question

Showing papers in "Computer Vision and Image Understanding in 2004"


Journal ArticleDOI
TL;DR: A new linear-time algorithm is presented in this paper that simultaneously labels connected components and their contours in binary images and extracts component contours and sequential orders of contour points, which can be useful for many applications.

599 citations


Journal ArticleDOI
TL;DR: A new cast shadow segmentation algorithm is proposed that exploits spectral and geometrical properties of shadows in a scene to perform this task and is robust and efficient in detecting shadows for a large class of scenes.

408 citations


Journal ArticleDOI
TL;DR: A new approach to high quality 3D object reconstruction, based on a deformable model, which defines the framework where texture and silhouette information can be fused by defining two external forces based on the images: a texture driven force and a silhouette driven force.

406 citations


Journal ArticleDOI
TL;DR: The use of layered probabilistic representations for modeling human activities is presented, and how the representation is used to do sensing, learning, and inference at multiple levels of temporal granularity and abstraction and from heterogeneous data sources is described.

370 citations


Journal ArticleDOI
TL;DR: A new representation and recognition method for human activities that recognizes multi-agent events by propagating the constraints and likelihood of event threads in a temporal logic network and presents results on real-world data and performance characterization on perturbed data.

351 citations


Journal ArticleDOI
TL;DR: Although the generalised color moment invariants are extracted from planar surface patches, it is argued that invariant neighbourhoods offer a concept through which they can also be used to deal with 3D objects and scenes.

279 citations


Journal ArticleDOI
TL;DR: Results obtained show that retrieval effectiveness increases in non-cascaded region-based querying by combined index, and also based on a combined color shape location index.

162 citations


Journal ArticleDOI
TL;DR: A new approach to the geometric alignment of a point cloud to a surface and to related registration problems which relies on instantaneous kinematics and on the geometry of the squared distance function of a surface is presented.

159 citations


Journal ArticleDOI
TL;DR: Experimental results with quantitative performance evaluations demonstrate the effectiveness of a PC cluster system for real-time reconstruction of dynamic 3D object action from multiview video images, a deformable 3D mesh model for reconstructing the accurate dynamic 2D object shape, and an algorithm of rendering natural-looking texture on the3D object surface from the multi-view video images.

144 citations


Journal ArticleDOI
TL;DR: Results show that the histograms made by GMVQ with a penalized log-likelihood (LL) distortion yield better retrieval performance for color images than the conventional methods of uniform quantization and VQ with squared error distortion.

131 citations


Journal ArticleDOI
TL;DR: The TSV transform provides an efficient way to remove noise by focusing on stable velocities, and constructs noise-free blobs, and is applied to tracking human figures in a sidewalk environment and extended to an interaction recognition system.

Journal ArticleDOI
TL;DR: An automatic road sign detection and recognition system that is based on a computational model of human visual recognition processing and the experimental results revealed both the feasibility of the proposed computational model and the robustness of the developedRoad sign detection system.

Journal ArticleDOI
TL;DR: This work proposes a new, simple and fast EDT in two scans using a 3 × 3 neighborhood, and develops an optimal two-scan algorithm to achieve the EDT correctly and efficiently in a constant time without iterations.

Journal ArticleDOI
TL;DR: The proposed angular optimization algorithms take advantage of adaptive stack filters design and weighted median filtering framework and are able to remove image noise, while maintaining excellent signal-detail preservation capabilities and sufficient robustness for a variety of signal and noise statistics.

Journal ArticleDOI
TL;DR: This paper proposes a method for helping to identify adult web sites by using the imagecontent as means of detecting erotic material, and proves to be quite successful in tests where all 20 sites where classified correctly.

Journal ArticleDOI
TL;DR: It is argued that biometric match score accuracy is best expressed in terms of a curve, the Receiver Operating Characteristic curve, and confidence intervals, or margins of error, should be provided for this curve for determining whether accuracy differences between systems are really statistically significant.

Journal ArticleDOI
TL;DR: 3-D rooftop boundary hypotheses are found from the line and junction features of the images by applying consecutive grouping procedures and are verified with evidence collected from the images and the elevation data.

Journal ArticleDOI
TL;DR: This paper presents an efficient image-based approach to navigate a scene based on only three wide-baseline uncalibrated images without the explicit use of a 3D model, and demonstrates three applications of the tri-view morphing algorithm.

Journal ArticleDOI
TL;DR: The core of the presented work describes the calibration and orientation of the images, mostly based on photogrammetric techniques, and the reconstruction of the 3-D body model in point cloud form.

Journal ArticleDOI
TL;DR: The possibility of an alternative model for motion perception based on synchronization with the transient oscillations of temporal band-pass filters that is consistent with other proposed models for human perception is discussed.

Journal ArticleDOI
TL;DR: The designed segmentation method can be extended to images for which it is required to segment a region of interest using an unsupervised approach, and is applied to Satellite Pour l'Observation de la Terre remote multispectral images.

Journal ArticleDOI
TL;DR: This paper defines a new dissimilarity measure that is more reliable than the Euclidean distance and yet computationally less expensive than EMD, and a mathematically sound definition of mean histogram can be defined for histogram clustering applications.

Journal ArticleDOI
TL;DR: This paper presents automatic image orientation detection algorithms based on both the luminance (structural) and chrominance (color) low-level content features based on the statistical learning support vector machines (SVMs) as the classifiers.

Journal ArticleDOI
TL;DR: This paper shows how to represent and extract interest points at variable scales and devise a method allowing the comparison of two images at two different resolutions, using a photometric- and rotation-invariant descriptors and an image matching strategy based on local constraints and on the robust estimation of this geometric model.

Journal ArticleDOI
TL;DR: The two-dimensional topological map is defined, a model which represents both topological and geometrical information of aTwo-dimensional labeled image which is minimal, complete, and unique and can be used to define efficient image processing algorithms.

Journal ArticleDOI
TL;DR: This paper proposes a novel system that is able to automatically detect and classify baseball highlights by seamlessly integrating image, audio, and speech clues using a unique framework based on maximum entropy model (MEM).

Journal ArticleDOI
TL;DR: A real time system for detecting repeated video clips from a live video source such as news broadcasts that utilizes customized temporal video segmentation techniques to automatically partition the digital video signal into semantically sensible shots and scenes.

Journal ArticleDOI
TL;DR: A framework for color image segmentation is presented, which combines color histogram analysis and region merging approach, and testing this algorithm with both artificially generated and real images shows quite reliable results.

Journal ArticleDOI
TL;DR: A framework for event detection is proposed where events, objects, and other semantic concepts are detected from video using trained classifiers and integration of content-based and concept-based querying in the search process is integrated.

Journal ArticleDOI
TL;DR: The experimental results showed that the present algorithm can cope with noisy figures, projective transformations, and complex occlusions.