Date/heure
Date(s) - 13 mars 2014

Catégories Pas de Catégories

$Automatic cinematography and film editing for 3-D animation\n\nBy Rémi Ronfard, INRIA Rhone Alpes.\n\nI will present some recent work on automatically creating movies from 3-D animation. This includes work on planning the positions and movements of cameras to capture the animation, and work on cutting between cameras (film editing or montage). I will present the prose storyboard language as a means for planning and directing cameras in 3-D animation. I will present ongoing work in analyzing the cinematography and editing of real movies using computer vision techniques. I will present recent work on implementing intelligent behaviors for autonomous cameras with high-level cinematographic goals. I will end with recent results on a computational model of film editing, where we analyze the main causes of editing errors and build a cost function for minimizing them. With a plausible semi-Markov assumption, we can solve for the best movie using dynamic programming. This new approach promises to significantly extend the expressiveness and naturalness of virtual movie-making.[$

Date/heure
Date(s) - 13 mars 2014

Catégories Pas de Catégories

$Automatic cinematography and film editing for 3-D animation\n\nBy Rémi Ronfard, INRIA Rhone Alpes.\n\nI will present some recent work on automatically creating movies from 3-D animation. This includes work on planning the positions and movements of cameras to capture the animation, and work on cutting between cameras (film editing or montage). I will present the prose storyboard language as a means for planning and directing cameras in 3-D animation. I will present ongoing work in analyzing the cinematography and editing of real movies using computer vision techniques. I will present recent work on implementing intelligent behaviors for autonomous cameras with high-level cinematographic goals. I will end with recent results on a computational model of film editing, where we analyze the main causes of editing errors and build a cost function for minimizing them. With a plausible semi-Markov assumption, we can solve for the best movie using dynamic programming. This new approach promises to significantly extend the expressiveness and naturalness of virtual movie-making.[$