Automatic Camera Control: A Dynamic Multi-Objective Perspective

Abstract

Abstract. Automatically generating computer animations is a challenging and complex problem with applications in games and film production. In this paper, we investigate how to translate a shot list for a virtual scene into a series of virtual camera configurations -i.e automatically controlling the virtual camera. We approach this problem by modelling it as a dynamic multi-objective optimisation problem and show how this metaphor allows a much richer expressiveness than a classical single objective approach. Finally, we showcase the application of a multi-objective evolutionary algorithm to generate a shot for a sample game replay and we analyse the results.

Virtual photography using multi-objective particle swarm optimization

“…Recently, Burelli and Preuss used multi-objective optimization for automatic composition [9]. There, a nondominated hypervolume is used as a fitness measure.…”

“…Abdullah et al [1] use a PSO to optimize cinematographic images via composition rules. Burelli and Preuss use multi-objective optimization for automatic cinematography [9]. This paper presents a new technique for virtual photography.…”

Virtual Camera Movement with Particle Swarm Optimization and Local Regression

“…For the dynamic scene, for example a cinematographic replay for a 3D shooter game, a multi-objective approach was proposed by Burelli and Preuss [5]. They prefered the S-Metric Selection Evolutionary Multiobjective Optimization Algorithm (SMS-EMOA) than NonDominated Sorting Algorithm-II because it is known to deal with a higher number of objectives.…”

“…A similar research has been attempted previously, although for different domains and goals. Virtual camera control can be applied to several applications like, airplane emergency simulation, to create cinematic replay of computer games [5] and [6], to create flythrough animation for 3D architectural models [7], to create stereoscopic photography in virtual environment [8], combined with UAV to for automatic cinematography and to capture cinema scene [9], [10], and [11]. Drucker et al [12] presented constraint-based for camera control in the virtual environment.…”

“…Ranon et al [23] used an improved version of PSO by tuning the parameter as in comparison with [17] works. Burelli et al [5] proposed a multi-objective approach for generating replay in the game. Burelli et al [24] proposed benchmark methods for virtual camera control.…”