20 likes | 158 Vues
This presentation by Mauricio Hess-Flores, Mark A. Duchaineau, and Kenneth I. Joy explores novel techniques in sequential reconstruction using parallax paths for feature tracking and structure updating. By analyzing intra-camera and inter-camera constraints, the study enhances outlier detection and corrections while improving scene structure and drift detection. Presented during the Video Analysis and Action Recognition session at the University of California, Davis, this work aims to advance methodologies in the field of data analysis and visualization, showcasing their implications for video processing and action recognition.
E N D
Click to edit title • Click to edit text Sequential Reconstruction Segment-Wise Feature Track and Structure Updating Based on Parallax Paths Mauricio Hess-Flores1, Mark A. Duchaineau2, and Kenneth I. Joy1 1 Institute for Data Analysis and Visualization, University of California, Davis, USA mhessf@ucdavis.edu, kenneth.i.joy@gmail.com 2 Lawrence Livermore National Laboratory, Livermore, CA, USA duchaine@google.com (Now at Google, Inc.) Session: Video Analysis and Action Recognition, Friday 9 November 2012
Mauricio Hess-Flores1, Mark A. Duchaineau2, and Kenneth I. Joy1 1Institute for Data Analysis and Visualization, University of California, Davis 2Lawrence Livermore National Laboratory (now at Google, Inc.) Sequential Reconstruction Segment-Wise Feature Track and Structure Updating Based on Parallax Paths Parallax paths concept and constraints: Parallax paths Intra-camera Inter-camera Concept Motivation Prediction Constraints Replicas (b) (c) (d) (a) Results: Feature track outlier detection and correction Improvement in scene structure Drift detection and correction