Publisher Description

Welcome to the 2008EuropeanConference onComputer Vision. These proce- ings are the result of a great deal of hard work by many people. To produce them, a total of 871 papers were reviewed. Forty were selected for oral pres- tation and 203 were selected for poster presentation, yielding acceptance rates of 4.6% for oral, 23.3% for poster, and 27.9% in total. Weappliedthreeprinciples.First,sincewehadastronggroupofAreaChairs, the ?nal decisions to accept or reject a paper rested with the Area Chair, who wouldbeinformedbyreviewsandcouldactonlyinconsensuswithanotherArea Chair. Second, we felt that authors were entitled to a summary that explained how the Area Chair reached a decision for a paper. Third, we were very careful to avoid con?icts of interest. Each paper was assigned to an Area Chair by the Program Chairs, and each Area Chair received a pool of about 25 papers. The Area Chairs then identi?ed and rankedappropriatereviewersfor eachpaper in their pool, and a constrained optimizationallocated three reviewers to each paper. We are very proud that every paper received at least three reviews. At this point, authors were able to respond to reviews. The Area Chairs then needed to reach a decision. We used a series of procedures to ensure careful review and to avoid con?icts of interest. ProgramChairs did not submit papers. The Area Chairs were divided into three groups so that no Area Chair in the group was in con?ict with any paper assigned to any Area Chair in the group.

Back Cover

The four-volume set comprising LNCS volumes 5302/5303/5304/5305 constitutes the refereed proceedings of the 10th European Conference on Computer Vision, ECCV 2008, held in Marseille, France, in October 2008. The 243 revised papers presented were carefully reviewed and selected from a total of 871 papers submitted. The four books cover the entire range of current issues in computer vision. The papers are organized in topical sections on recognition, stereo, people and face recognition, object tracking, matching, learning and features, MRFs, segmentation, computational photography and active reconstruction.

Table of Contents

Lecture by Prof. Jan Koenderink.- Something Old, Something New, Something Borrowed, Something Blue.- Recognition.- Learning to Localize Objects with Structured Output Regression.- Beyond Nouns: Exploiting Prepositions and Comparative Adjectives for Learning Visual Classifiers.- Learning Spatial Context: Using Stuff to Find Things.- Segmentation and Recognition Using Structure from Motion Point Clouds.- Poster Session I.- Keypoint Signatures for Fast Learning and Recognition.- Active Matching.- Towards Scalable Dataset Construction: An Active Learning Approach.- GeoS: Geodesic Image Segmentation.- Simultaneous Motion Detection and Background Reconstruction with a Mixed-State Conditional Markov Random Field.- Semidefinite Programming Heuristics for Surface Reconstruction Ambiguities.- Robust Optimal Pose Estimation.- Learning to Recognize Activities from the Wrong View Point.- Joint Parametric and Non-parametric Curve Evolution for Medical Image Segmentation.- Localizing Objects with Smart Dictionaries.- Weakly Supervised Object Localization with Stable Segmentations.- A Perceptual Comparison of Distance Measures for Color Constancy Algorithms.- Scale Invariant Action Recognition Using Compound Features Mined from Dense Spatio-temporal Corners.- Semi-supervised On-Line Boosting for Robust Tracking.- Reformulating and Optimizing the Mumford-Shah Functional on a Graph — A Faster, Lower Energy Solution.- Viewpoint Invariant Pedestrian Recognition with an Ensemble of Localized Features.- Perspective Nonrigid Shape and Motion Recovery.- Shadows in Three-Source Photometric Stereo.- Hamming Embedding and Weak Geometric Consistency for Large Scale Image Search.- Estimating Geo-temporal Location of Stationary Cameras Using Shadow Trajectories.- An Experimental Comparison of Discrete and Continuous Shape Optimization Methods.- Image Feature Extraction Using Gradient Local Auto-Correlations.- Analysis of Building Textures for Reconstructing Partially Occluded Facades.- NonrigidImage Registration Using Dynamic Higher-Order MRF Model.- Tracking of Abrupt Motion Using Wang-Landau Monte Carlo Estimation.- Surface Visibility Probabilities in 3D Cluttered Scenes.- A Generative Shape Regularization Model for Robust Face Alignment.- Modeling and Recognition of Landmark Image Collections Using Iconic Scene Graphs.- VideoCut: Removing Irrelevant Frames by Discovering the Object of Interest.- ASN: Image Keypoint Detection from Adaptive Shape Neighborhood.- Online Sparse Matrix Gaussian Process Regression and Vision Applications.- Multi-stage Contour Based Detection of Deformable Objects.- Brain Hallucination.- Range Flow for Varying Illumination.- Some Objects Are More Equal Than Others: Measuring and Predicting Importance.- Robust Multiple Structures Estimation with J-Linkage.- Human Activity Recognition with Metric Learning.- Shape Matching by Segmentation Averaging.- Search Space Reduction for MRF Stereo.- Estimating 3D Face Model and Facial Deformation from a Single Image Based on Expression Manifold Optimization.- 3D Face Recognition by Local Shape Difference Boosting.- Efficiently Learning Random Fields for Stereo Vision with Sparse Message Passing.- Recovering Light Directions and Camera Poses from a Single Sphere.- Tracking with Dynamic Hidden-State Shape Models.- Interactive Tracking of 2D Generic Objects with Spacetime Optimization.- A Segmentation Based Variational Model for Accurate Optical Flow Estimation.- Similarity Features for Facial Event Analysis.- Building a Compact Relevant Sample Coverage for Relevance Feedback in Content-Based Image Retrieval.- Discriminative Learning for Deformable Shape Segmentation: A Comparative Study.- Discriminative Locality Alignment.- Stereo.- Efficient Dense Scene Flow from Sparse or Dense Stereo Data.- Integration of Multiview Stereo and Silhouettes Via Convex Functionals on Convex Domains.- Using Multiple Hypotheses to Improve Depth-Maps for Multi-View Stereo.- Sparse Structures in L-Infinity Norm Minimization for Structure and Motion Reconstruction.

Long Description

Welcome to the 2008EuropeanConference onComputer Vision. These proce- ings are the result of a great deal of hard work by many people. To produce them, a total of 871 papers were reviewed. Forty were selected for oral pres- tation and 203 were selected for poster presentation, yielding acceptance rates of 4.6% for oral, 23.3% for poster, and 27.9% in total. Weappliedthreeprinciples.First,sincewehadastronggroupofAreaChairs, the ?nal decisions to accept or reject a paper rested with the Area Chair, who wouldbeinformedbyreviewsandcouldactonlyinconsensuswithanotherArea Chair. Second, we felt that authors were entitled to a summary that explained how the Area Chair reached a decision for a paper. Third, we were very careful to avoid con'icts of interest. Each paper was assigned to an Area Chair by the Program Chairs, and each Area Chair received a pool of about 25 papers. The Area Chairs then identi'ed and rankedappropriatereviewersfor eachpaper in their pool, and a constrained optimization allocated three reviewers to each paper. We are very proud that every paper received at least three reviews. At this point, authors were able to respond to reviews. The Area Chairs then needed to reach a decision. We used a series of procedures to ensure careful review and to avoid con'icts of interest. ProgramChairs did not submit papers. The Area Chairs were divided into three groups so that no Area Chair in the group was in con'ict with any paper assigned to any Area Chair in the group.

Details