A Benchmark for Virtual Camera Control

Abstract

Abstract. Automatically animating and placing the virtual camera in a dynamic environment is a challenging task. The camera is expected to maximise and maintain a set of properties -i.e. visual composition -while smoothly moving through the environment and avoiding obstacles. A large number of different solutions to the problem have been proposed so far including, for instance, evolutionary techniques, swarm intelligence or ad hoc solutions. However, the large diversity of the solutions and the lack of a common benchmark, made any comparative analysis of the different solutions extremely difficult. For this reason, in this paper, we propose a benchmark for the problem of virtual camera control and we analyse a number of different problems in different virtual environments. Each of these scenarios is described through a set of complexity measures and, as a result of this analysis, a subset of scenarios is selected as the core of the benchmark.

Virtual Camera Movement with Particle Swarm Optimization and Local Regression

“…A benchmark for virtual camera control was proposed by Burelli and Yannakakis [24] by measuring the accuracy (how close the best camera configuration is found by the algorithm), reliability (how often the algorithm succeeds to provide an acceptable solution), and initial convergence time (how much time it takes the algorithm to converge to an optimal camera configuration). Three test environment (a forest, a house, and a rocky clearing) were selected to evaluate levels of complexity.…”

“…As in [24], vantage angles (1), object projection size (2) and object visibility (3) as the satisfaction function can be defined respectively as follows: (1) where is the desired horizontal angle between camera and target's front vector, is the desired vertical angle between camera and target's front vector, and is the normalized camera position transformed in targeted relative space: (2) where is the desired projection size, and are the screen height and screen width respectively and and are the projected height and width of the target object's proxy geometry respectively:…”

“…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.…”

Profiling Director’s Style Based on Camera Positioning Using Fuzzy Logic

“…A benchmark for virtual camera control is proposed [38] by measuring the accuracy, reliability, and initial convergence time. The simulation uses three scene backgrounds (forest, house and rocky).…”