CONTENTS -- Preface -- 1. An appearance-based method for parametric video registration X. Orriols, L. Barceló and X. Binefa -- 1.1. Introduction -- 1.2. Appearance Based Framework for Multi-Frame Registration -- 1.2.1. Appearance Representation Model -- 1.2.2. Polynomial Surface Model -- 1.2.3. The Algorithm -- 1.3. Experimental Results -- 1.3.1. Selecting a Reference Frame. Consequences in the Registration -- 1.3.2. Analyzing the Complexity in the Polynomial Model. Towards 3D Affine Reconstruction -- 1.4. Summary and Conclusions -- Acknowledgments -- References -- 2. An interactive algorithm for image smoothing and segmentation M. C. de Andrade -- 1. Introduction -- 2. The interactive image smoothing and segmentation algorithm - ISS -- 2.1. Edge preserving smoothing under controlled curvature motion -- 2.1.1. Stopping criteria for curvature based denoising -- 2.1.2. Effect of denoising on the ISS -- 2.2. The interactive region growing and merging step -- 2.3. The ISS algorithm steps -- 3. Applications -- 4. Conclusions and Outlook -- Acknowledgments -- Appendix A. ISS Pseudo-code -- Appendix B. ISS Execution time for known test-images -- References -- 3. Relevance of multifractal textures in static images A. Turiel -- 3.1. Introduction -- 3.2. Multifractal framework -- 3.3. Multifractal decomposition -- 3.4. Reconstructing from edges -- 3.5. Relevance of the fractal manifolds -- 3.6. Conclusions -- Acknowledgements -- References -- 4. Potential fields as an external force and algorithmic improvements in deformable models A. Caro et al. -- 4.1. Introduction -- 4.1.1. Overview on Active Contours -- 4.1.2. Scope and purpose of the research -- 4.2. Algorithm Design -- 4.2.1. Standard Deformable Models -- 4.2.2. The new approach for Deformable Models -- 4.2.3. A practical application: DeformableModels on Iberian ham MRI.

4.3. Practical Results and their Discussion -- 4.4. Conclusions -- Acknowledgements -- References -- 5. Optimization of weights in a multiple classifier handwritten word recognition system using a genetic algorithm S. Günter and H. Bunke -- 5.1. Introduction -- 5.2. Handwritten word recognizer -- 5.2.1. Preprocessing -- 5.2.2. Feature extraction -- 5.2.3. Hidden Markov models -- 5.3. Ensemble creation methods -- 5.3.1. Issues in ensemble creation -- 5.3.2. Bagging -- 5.3.3. AdaBoost -- 5.3.4. Random subspace method -- 5.3.5. Architecture variation -- 5.4. Combination schemes -- 5.4.1. Maximum score rule -- 5.4.2. Performance weighted voting -- 5.4.3. Weighted voting using weights calculated by a genetic algorithm -- 5.4.4. Voting with ties handling -- 5.5. Genetic algorithm for the calculation of the weights used by weighted voting -- 5.5.1. Chromosome representation and fitness -- 5.5.2. Initialization and termination -- 5.5.3. Crossover operator -- 5.5.4. Mutation operator -- 5.5.5. Generation of a new population -- 5.6. Experiments -- 5.7. Conclusions -- Acknowledgments -- Appendix A. HandwrittenWord Samples -- References -- 6. Dempster-Shafer's basic probability assignment based on fuzzy membership functions A. O. Boudraa et al. -- 6.1. Introduction -- 6.2. Dempster-Shafer theory -- 6.3. Fuzzy approach -- 6.4. Basic probability assignment -- 6.5. Results -- 6.6. Conclusion -- References -- 7. Automatic instrument localization in laparoscopic surgery J. Climent and P. Mars -- 7.1. Introduction -- 7.2. System Description -- 7.2.1. Filtering stage -- 7.2.2. Edge orientation extraction -- 7.2.3. Hough transform computation -- 7.2.4. Segment extraction -- 7.2.5. Heuristic filter -- 7.2.6. Position prediction -- 7.2.7. Target selection -- 7.3. Results -- 7.4. Conclusion -- Acknowledgements -- References.

8. A fast fractal image compression method based on entropy M. Hassaballah, M. M. Makky and Y. B. Mahdy -- 1. Introduction -- 2. Fractal Image Coding -- 2.1. Principle of Fractal Coding -- 2.2. Baseline Fractal Image Coding Algorithm -- 3. The Proposed Method -- 3.1. Entropy -- 3.2. The Entropy Based Encoded Algorithm -- 4. Experimental Results -- 5. Conclusions -- References -- 9. Robustness of a blind image watermark detector designed by orthogonal projection C. Jin and J. Peng -- 1. Introduction -- 2. The Design Method of the Watermark Detector -- 3. Experiment Results and Discussion -- 3.1. Performance Test of Two kinds of Methods -- 3.2. Test of Anti-Noise Attack -- 3.3. Test of Anti-Rotation Attack -- 3.4. Test of Anti-Translation Attack -- 3.5. Test of Anti-Other Attack -- 4. Conclusion -- References -- 10. Self-supervised adaptation for on-line script text recognition L. Prevost and L. Oudot -- 10.1. Introduction -- 10.2. Literature review -- 10.3. Writer independent baseline system -- 10.4. Writer adaptation strategies -- 10.4.1. Supervised adaptation -- 10.4.2. Self-supervised adaptation -- 10.4.2.1. Systematical activation (SA) -- 10.4.2.2. Conditional activation (CA) -- 10.4.2.3. Dynamic management (DM) -- 10.5. Supervised / self-supervised combination -- 10.6. Conclusions & Future works -- References -- 11. Combining model-based and discriminative approaches in a modular two-stage classification system: Application to isolated handwritten digit recognition J. Milgram, R. Sabourin and M. Cheriet -- 1. Introduction -- 2. Model-based approach -- 2.1. Characterization of the pattern recognition problem -- 2.2. Modeling data with hyperplanes -- 2.3. Estimate posterior probability -- 3. Combination with discriminative approach -- 3.1. Conflict detection -- 3.2. Use of Support Vector Classifiers -- 3.3. Re-estimate posterior probabilities.

4. Experimental results -- 4.1. Model-based approach -- 4.2. Support Vector Classifiers -- 4.3. Two-stage classification system -- 5. Conclusions and perspectives -- References -- 12. Learning model structure from data: An application to on-line handwriting H. Binsztok and T. Arti`eres -- 12.1. Introduction -- 12.2. Building an initial HMM from training data -- 12.2.1. Building a left-right HMM from a training sequence -- 12.2.1.1. HMM structure -- 12.2.1.2. Parameters -- 12.2.2. Building the initial global HMM -- 12.3. Iterative simplification algorithm -- 12.4. Using the approach for clustering and for classification -- 12.5. Experimental databases -- 12.5.1. Artificial data -- 12.5.2. On-line handwritten signals -- 12.6. Probability density function estimation -- 12.6.1. Artificial data -- 12.6.2. Handwritten signals -- 12.7. Clustering experiments -- 12.7.1. Evaluation criteria -- 12.7.2. Benchmark method -- 12.7.3. Experiments on artificial data -- 12.7.4. Experiments on on-line handwritten signals -- 12.8. Classification experiments -- 12.9. Conclusion -- References -- 13. Simultaneous and causal appearance learning and tracking J. Melenchón, I. Iriondo and L. Meler -- 13.1. Introduction -- 13.2. Incremental SVD with Mean Update -- 13.2.1. Fundamentals -- 13.2.2. Updating SVD -- 13.2.3. Updating SVD and Mean -- 13.2.4. Mean Extraction from a Given SVD -- 13.2.5. Time and Memory Complexity -- 13.3. On-the-Fly Face Training -- 13.3.1. Data Representation -- 13.3.2. Training Process -- 13.3.3. Cost Analysis -- 13.4. Experimental Results -- 13.4.1. On-the-Fly Training Algorithm -- 13.4.2. Incremental SVD and Mean Computation -- 13.4.2.1. Precision comparisons -- 13.4.2.2. Execution time -- 13.4.2.3. Conclusions -- 13.5. Concluding Remarks -- References -- 14. A comparison framework for walking performances using aSpaces J. Gonz`alez et al.

14.1. Introduction -- 14.2. RelatedWork -- 14.3. Defining the Training Samples -- 14.2. RelatedWork -- 14.3. Defining the Training Samples -- 14.4. The aWalk aSpace -- 14.5. Parametric Action Representation: the p-action -- 14.6. Human Performance Comparison -- 14.7. Arc length Parameterization of p-actions -- 14.8. Experimental Results -- 14.9. Conclusions and FutureWork -- Acknowledgements -- References -- 15. Detecting human heads with their orientations A. Sugimoto, M. Kimura and T. Matsuyama -- 15.1. Introduction -- 15.2. Contour model for human-head appearances -- 15.2.1. Human head and its appearances -- 15.2.2. Evaluation of contour model -- 15.3. Inner models for face orientations -- 15.3.1. Facial components -- 15.3.2. Detecting facial components using Gabor-Wavelets -- 15.3.3. Inner models of head appearances with facial components -- 15.4. Algorithm -- 15.5. Experimental evaluation -- 15.5.1. Evaluation on face orientations using a face-image database -- 15.5.2. Evaluation in the real situation -- 15.5.2.1. Human-head detection in the real situation -- 15.5.2.2. Effectiveness of human-head evaluation -- 15.6. Conclusion -- Acknowledgements -- References -- 16. Prior knowledge based motion model representation A. D. Sappa et al. -- 16.1. Introduction -- 16.2. PreviousWorks -- 16.3. The Proposed Approach -- 16.3.1. Body Modeling -- 16.3.2. Feature Point Selection and Tracking -- 16.3.2.1. Feature Point Selection -- 16.3.2.2. Feature Point Tracking -- 16.3.3. Motion Model Tuning -- 16.4. Experimental Results -- 16.5. Conclusions and FutureWork -- Acknowledgements -- References -- 17. Combining particle filter and population-basedmetaheuristics for visual articulated motion tracking J. J. Pantrigo et al. -- 17.1. Introduction -- 17.2. Particle Filters -- 17.3. Population-BasedMetaheuristics -- 17.3.1. Path Relinking -- 17.3.2. Scatter Search.

17.4. Particle Filter and Population-BasedMetaheuristics Hybrid Algorithms.